Overview

Plan
Microsoft
Education
AI Toolkit

Implement
A navigator
for education
institutions
to plan their

AI Navigators
AI journey
Research

April 2024 | Version 1.5

Welcome to the

Overview
Microsoft Education AI Toolkit

At Microsoft, we stand ready to support you as our advancements in AI are grounded in our mission
to empower every person and every organization on the planet to achieve more. Generative AI
technologies like Microsoft Copilot are changing the way we research, work, and learn— and we
share your excitement in how they are already being used. This toolkit has been specifically created
for education leaders to provide knowledge, strategies, and recommendations about their effective
and responsible use so you can begin your AI journey today.

Plan
We are committed to creating technologies that are accessible, inclusive, and tailored to meet
the diverse needs of all learners. Our AI systems are designed responsibly—keeping people at
the center of safe, secure, and trustworthy use of these tools.

In the pages that follow, you’ll be introduced to a variety of technologies including Microsoft Copilot,
Copilot for Microsoft 365, Copilot in Windows, GitHub Copilot, and Azure AI Studio as well as the
stories and best practices that showcase how they are already being used by education institutions
across the globe–along with the latest research that demonstrates the positive outcomes these AI
solutions are having. We’ve also provided step-by-step instructions, screenshots, and links so you

Implement
and your team can try these amazing tools for yourself.

As we continue to advance these technologies, we recognize the important role organizations like
yours will play in shaping the future of teaching using these new tools. Your engagement with this
resource will deepen your understanding of generative AI and will provide a means to learn from the
experiences of other educators and institutions–those we call AI Navigators.

Thank you for your interest in Microsoft’s generative AI technologies and all you do to prepare
the next generation of leaders and innovators. We look forward to continuing this exciting
journey with you.

AI Navigators
Yours sincerely,

Paige Johnson
Paige Johnson
VP of Global Education
and Media Industry Marketing
Microsoft Corporation
Research
Table of Contents

Overview
Section 1 Overview 2
Welcome letter 2

How to use this resource 4

Using Microsoft Copilot 4

Plan
Introduction to AI 5

Section 2 Plan 18
Leveraging responsible AI frameworks 20

Institutional policy considerations 22

The importance of a secure foundation 28

Implement
Data governance 39

Engaging your community 42

Section 3 Implement 50
Getting started with Microsoft AI 51

Microsoft Copilot capabilities 72

Creating effective prompts with Microsoft Copilot 75

AI Navigators
Professional learning 84

Section 4 AI Navigators 87
Education AI navigators 88

Customer stories 90

Technical implementation guides 99

Section 5 Research 114

Research

Latest research findings 115

Frequently asked questions 133

How to use this resource

Overview
The Microsoft Education AI Toolkit is intended for education leaders—at universities, schools, state
departments, and ministries of education. The toolkit aims to equip educational leaders at various
stages of their AI journey with knowledge, strategies, and tips to confidently advance their use of
generative AI technologies.

Organized into five main categories—Overview, Plan, Implement, AI Navigators, and Research—you can
use the navigation tabs on the right-hand side of the PDF to explore frameworks, guidelines, examples
and much more.

Plan
Using Microsoft Copilot
Microsoft Copilot is your everyday AI assistant. There are several ways to access Copilot including any
modern web browser, on your computer, and even on your mobile devices as a standalone application.
For education customers, Microsoft Copilot is free to use with your Microsoft login. When you use your
academic credentials, you'll have access to data and copyright protection. As you read through this
guide, you’ll encounter pre-scripted AI prompts which, when selected, will open so you can experience
the tool for yourself. Depending on your default browser, you may have to select enter for the
pre-scripted prompt like the one below to activate and return a result.

Implement
Microsoft Copilot provides three conversation styles, or modes: More Creative, More Balanced, and
More Precise. When experimenting with the prompts in this toolkit, try the different conversation styles
to tailor your results.

Get interactive: Explore AI prompts firsthand

Click anywhere on the prompt below to see how this toolkit feature works.

AI Navigators
Copilot prompt

Assume the role of an education institution leader such as a provost, dean, or

superintendent for a medium-sized institution with over 10,000 students and provide
a set of five guiding questions and summary responses for my education institution to
consider ensuring the responsible use of generative AI.
Research

Navigate back to
4 Microsoft Education AI Toolkit
table of contents
Introduction to AI

Overview
With the release of ChatGPT in late 2022, generative AI has surfaced as one of the most talked
about technologies since the emergence of the internet in the mid-1990’s. It has transformed the
technological landscape and is already impacting the worlds of work, education, and entertainment. At
its core, generative AI leverages complex algorithms and vast datasets to generate new, original content
ranging from text, images, music, video, and more.

Grasping foundational concepts of AI technology is essential for your role, guiding the decision-
making process on the best implementation methods to reach your goals and ensure a more equitable
future for students. As an IT and education leader, you are at the forefront of integrating technological
advancements not only into educational practices but also into the broader operational framework

Plan
of your organization, which includes enhancing digital security, ensuring data privacy, managing
infrastructure, and fostering collaborative learning environments. This holistic approach supports a
seamless and safe educational journey for all students.

1950s
Artificial Intelligence Artificial Intelligence
The field of computer science that seeks to create intelligent
machines that can replace or exceed human intelligence.

Implement
1959
Machine Learning Machine Learning
Subset of AI that enables machines to learn from existing data
and improve upon that data to make decisions or predictions.

2017
Deep Learning Deep Learning
A machine learning technique in which layers of neural
networks are used to process data and make decisions.

2021

AI Navigators
Generative AI
Create new written, visual, and auditory content given
Generative AI prompts or existing data.

Research

Navigate back to
5 Microsoft Education AI Toolkit
table of contents
 Overview
Copilot prompt

You are a computer scientist who works with AI. Explain the growth of AI to an audience of
K-20 IT professionals and school leaders. Give clear and easy-to-understand explanation of AI,
demystify AI and inspire innovative educational applications. Then give 5 unique examples of
how AI is currently used in educational settings from personalizing learning to helping with
administrative efficiency.

Plan
The subsequent pages in the Overview section offer an array of practical and contextualized insights.

• Explore AI-related foundational vocabulary in Terms.

• Navigate through a concise evolution of AI technology in A brief overview of AI.

• Delve into data’s central role in education, by reading It’s all about the data.

Implement
• Examine AI's Impact on work skills in AI and the future of work.

• Scan the functions of each copilot in Get to know the Microsoft Copilot tools.

• Explore suggestions for how different educational practitioners might apply a copilot in Copilot for
IT leaders, Copilot for education leaders, and Copilot for educators.

• Plan for student interaction with AI by reading AI for students.

AI Navigators
• Meet the AI-powered tools that boost student learning in Learning Accelerators.

• Engage with the sample copilot prompts sprinkled throughout the section.
Research

Navigate back to
6 Microsoft Education AI Toolkit
table of contents
Terms

Overview
Algorithm
A set of clear and specific instructions that can be performed in a prescribed sequence to achieve a particular goal and
that has a recognizable set of end conditions.

Artificial Intelligence (AI)

Defined as “the ability of a computer or other machine to perform those activities [tasks] that are normally thought
to require intelligence.” AI tasks involve various data analyses or production such as providing predictions or
recommendations, language translation, computer vision systems, or speech recognition. AI is a human endeavor that
combines information about people and the physical world into mathematical constructs. Such technologies typically rely
on statistical methods, with the possibility for errors throughout an AI system’s lifespan.

Plan
Deep Learning
A machine learning technique in which layers of neural networks are used to process data and make decisions.

Generative AI (genAI)
A term for AI systems that generate various forms of novel output, including text, code, graphics, or audio. Examples
of generative AI include generative pre-trained transformer (GPT) chatbots and text-to-image generators.

Fabrication
A phenomenon of large language models (LLMs) sometimes generating responses that are factually incorrect

Implement
or incoherent.

Large Language Model (LLM)

A large language model (LLM) is a type of AI that can process and produce natural language text. It learns from a
massive amount of data gathered from sources like books, articles, webpages, and images to discover patterns and
rules of language.

Machine Learning (ML) model

Machine learning (ML) models typically involve data, code, and model outputs, while AI systems have other socio-
technical components, such as user interfaces. An ML model is trained to recognize certain types of patterns and then
uses an algorithm to make predictions about new data.

AI Navigators
Natural Language Processing (NLP)
NLP is the ability of a computer program to understand human language as it is spoken and written—it is a type of
artificial intelligence.

Neural Network
A machine learning model that uses algorithms to mimic the human brain
.
Training
A term that refers to providing a machine learning model’s algorithm with a given dataset for processing and
identifying patterns that the model will then use for performing predictive tasks in its deployment setting.
Research

Navigate back to
7 Microsoft Education AI Toolkit
table of contents
A brief overview of AI

Overview
Artificial Intelligence emerged in the 1950s when early pioneers like Alan Turing laid the groundwork for
machine intelligence, posing the question of whether machines could exhibit human-like thinking. In
1956, researchers met at the Dartmouth Conference to explore the idea of creating machines that could
think like humans—this is considered the beginning of AI.

Throughout the following decades, AI garnered periods of enthusiasm and skepticism. However,
as computers have become more powerful, and we have more data and better algorithms, AI has
advanced, especially the subset of AI called machine learning (ML). Neural networks and deep learning
techniques have made ML more powerful and useful. Advancements in Natural Language Processing
(NLP) have enabled machines to understand, interpret, and generate human language making AI even
more accessible to a diverse group of users.

Plan
Today, AI is used in many parts of our lives, from virtual assistants and recommendation systems
to autonomous vehicles and diagnostic tools. Smart home devices, email filtering, and language
translation apps also use AI. An area of significant advancement in AI technology is generative AI.
Generative AI allows any user to prompt the tool to create text, images, code snippets and more. What
sets generative AI apart is that it makes creating easy for nearly everyone, even if they don’t have
special technical skills. It offers a simple way for all users to make their own content.

It’s all about the data

Implement
Data is, in many ways, at the center of educational institutions, playing a pivotal role in shaping
strategies, enhancing teaching methodologies, and fostering continuous improvement. Starting the
AI journey in education requires a fundamental shift in perspective: Recognizing that thoughtful
consideration of data management is central to the strategic planning needed to leverage the full
potential of AI technologies. This isn't just about technology; it's about shaping a future where strategic
decisions are informed and enhanced by AI.

One of the more challenging barriers to implementing AI solutions in education is the existence of data

AI Navigators
silos, separate information repositories confined within various proprietary software systems. These
silos not only hinder access to data but also impede the holistic analysis necessary for data-driven
decision-making.

Addressing this challenge requires a concerted effort to dismantle data silos and embrace a more
inclusive approach to data aggregation. By breaking down these barriers and adopting a unified data
approach, schools can achieve deeper insights and offer personalized learning experiences, enhancing AI's
impact. This not only enhances the capability of AI systems to generate meaningful outcomes but also
significantly advances the institution's ability to meet the evolving needs of its students and educators
Research

Navigate back to
8 Microsoft Education AI Toolkit
table of contents
Initially, the focus should be on establishing a foundational data management strategy that begins
to break down silos and integrate diverse data types. Simultaneously, implementing basic security

Overview
measures to protect this data is essential. As institutions evolve, so too can their data management and
security practices, evolving from simple, initial setups to more sophisticated systems like data lakes and
advanced encryption methods.

This approach emphasizes the importance of starting where you are, with what you have, and
understanding that perfection is not a prerequisite for progress. Incremental improvements, fueled by
ongoing learning and adaptation, are key to building a robust framework capable of supporting the
sophisticated needs of AI technologies. Encouraging a culture of continuous improvement and shared
responsibility among all stakeholders—administrators, IT staff, faculty, and students—can significantly
enhance both data management practices and security postures.

Plan
Microsoft’s integrated cloud platform, Microsoft Azure, offers a comprehensive solution. Azure
simplifies the development of advanced solutions and the use of analytics. Educational institutions
using Microsoft cloud solutions can centralize their data management, increase operational flexibility,
and provide personalized experiences for students. This solution isn't just about technology, it's about
shaping a future where strategic decisions are informed and enhanced by AI. As a result, data driven
decisions informed by AI constitute a fundamental shift from storing, accessing, and analyzing data to
dialoguing and conversing with your data.

Quality and diversity of data over volume

Implement
The true power of AI is unlocked not by the sheer volume of data but by its diversity. Educational institutions
have access to a wide range of data types, including academic records, multimedia, and behavioral metrics.
However, the integration of these varied data sources is crucial to harness their full potential. For AI
systems to function well, they require a broad spectrum of data types. The key to unlocking valuable
insights from AI lies in the variety and volume of data. This diversity enables AI to generate more
comprehensive insights. The concept of big data is key here—it's not just about large amounts of
data, but about varied and comprehensive datasets that feed into large language models, significantly
enhancing their performance.

AI Navigators
The journey toward effective AI integration in education is marked by learning from challenges,
celebrating advancements, and remaining adaptable to the ever-changing digital landscape. Each
step forward, no matter how small, contributes to the overarching goal of transforming educational
environments through the power of AI, underpinned by solid data management and security foundations.
Research

Navigate back to
9 Microsoft Education AI Toolkit
table of contents
 AI and the future of work

Overview
Artificial Intelligence, including generative AI, is significantly
transforming the landscape of work, necessitating a reevaluation
of the skills required in the workforce. As AI integrates into various
industries, it creates new opportunities that require a blend of
technical proficiency and durable skills like critical thinking and
emotional intelligence. Recognizing these shifts, reports from
professional organizations underscore the urgency for an updated
skills framework, highlighting a growing demand for a workforce
adept at navigating a technology-driven environment.
“If our era is the next
Industrial Revolution, This evolution in the job market calls for a corresponding shift in

Plan
educational models. Traditional curricula, often criticized for their
as many claim, AI slow adaptation to technological advancements, must transition
is surely one of its towards more dynamic, personalized learning experiences that
foster technical literacy and cultivate the skills essential for the AI
driving forces.” era such as curiosity and meta-cognition. These skills are essential
for understanding AI and harnessing its potential. Proficiency in
prompt design is also crucial for effective information retrieval and
Dr. Fei-Fei Li
content creation.
Stanford University

Implement
The path forward involves strategic planning that integrates
AI and future skills into curricula, fostering partnerships with
technology companies and creating a culture that values
innovation and adaptability.

Our evolving educational landscape emphasizes the shift from

creating content to analyzing and integrating skills, driven by
AI’s capacity to rapidly locate and generate content. However,
it’s important to recognize that generative AI isn’t perfect and
can sometimes produce inaccurate information or fabrications.

AI Navigators
Students must be adept at identifying errors and inaccuracies.

To succeed in an AI collaborative world, students, educators,

perhaps every member of an educational community must adapt
or learn new skills like how to prioritize, delegate, proof, review,
and master efficiency. Equipping students and educators with these
skills is essential for success in an AI collaborative world.
Research

Navigate back to
10 Microsoft Education AI Toolkit
table of contents
Get to know the Microsoft Copilot tools

Overview
Microsoft Copilot Copilot for Microsoft 365

AI-powered chat assistant designed to AI-powered productivity tool that

aid users in web browsing and more. includes access to Microsoft Copilot
Commercial data protection is included and integrations with Copilot in
for education institutions. Word, PowerPoint, Excel, Outlook,

Plan
Teams, Loop, and other Microsoft
AI chat for the web with 365 applications.
commercial data protection
Learn more about Microsoft Copilot Works alongside you in the
applications you use every day
Learn more about
Copilot for Microsoft 365

Implement

Copilot in Windows

Copilot in Windows is an AI assistant

in Windows 11 that can help you
with various tasks, such as changing Copilot for Security
settings, organizing windows, getting
answers, and generating images. Security-focused generative AI
solution enhancing defender
A powerful combination of AI efficiency and capabilities.

AI Navigators
and productivity Using natural language assistive
Learn more about experience in various scenarios,
Copilot in Windows including incident response, threat
hunting, intelligence gathering, and
posture management.

Defend at machine
speed with Microsoft Copilot
for Security
Learn more about
Copilot for Security
Research

Navigate back to
table of contents
Get to know the Microsoft Copilot tools

Overview
Copilot for Dynamics 365 Copilot for Power Platform

A tool that helps organizations Microsoft Copilot for Power Platform

automate tasks, analyze data, and is a tool that helps educational users
give suggestions to improve school create and customize apps, workflows,
performance and student outcomes. and chatbots for their schools. It uses
artificial intelligence to generate code,
Turbocharge your staff with a suggestions, and solutions based on

Plan
copilot for every job role natural language inputs.
Learn more about
Copilot for Dynamics 365 Imagine it, describe it, and
Power Platform builds it
Learn more about
Copilot for Power Platform

Implement
GitHub Copilot Commercial data protection
GitHub Copilot is a coding assistant Safeguards user and organizational
that helps you write code faster and data by preventing prompts and
smarter by generating suggestions responses from being saved when
based on your context and description. signed in with a school-issued
Microsoft account.
Increase developer productivity
to accelerate innovation Learn more about

AI Navigators
Learn more about GitHub Copilot Commercial data protection

Customer Copyright Commitment

Customer Copyright Commitment
Microsoft defends customers from
IP infringement claims arising from
the customer’s use and distribution
of the output content generated by
Microsoft’s Copilot services.

Learn more about

Research

Customer Copyright Commitment

Navigate back to
table of contents
Copilot for IT leaders

Overview
IT leaders play a pivotal role in maintaining infrastructure assets, establishing cybersecurity protocols,
protecting private data, and supporting community members with technical assistance. Microsoft
Copilot provide ways to simplify and streamline these challenging responsibilities in schools and higher
education institutions.

Microsoft Copilot Copilot for Microsoft 365 Copilot for Security

Increase productivity and Complete specialized tasks Respond to external threats

Plan
save time performing that use Microsoft 365 apps and evaluate risks using
common IT duties like: and files like: natural language queries
and prompts designed to:
• Update acceptable use • Analyze device inventory
policy (AUP) spreadsheets • Assess incident impact

• Create FAQs for adopted • Translate ticket languages • Develop remediation plans
technologies
• Summarize IT candidate • Analyze vulnerabilities
• Draft step-by-step

Implement
resumes
tutorials

Copilot prompt
Open your institution’s Acceptable Use Policy (AUP) in the Edge browser. Open Copilot sidebar

AI Navigators
from the top right and enter the prompt below:

Please review the Information Technology Acceptable Use Policy on the page for potential
improvements. Specifically, look for any outdated information, areas in need of clarification,
inconsistencies in language, and suggestions for enhancing user understanding. Check for the
inclusion of the last update date, ensure accessibility considerations, and provide insights on the scope,
monitoring procedures, and contact information. Additionally, analyze the clarity of prohibitions,
suggest examples where helpful, and assess the completeness of related sections such as exceptions
and definitions. Your feedback should help identify any potential revisions to improve the overall
effectiveness, clarity, and user-friendliness of the policy.
Research

Navigate back to
13 Microsoft Education AI Toolkit
table of contents
Copilot for education leaders

Overview
The role of an education leader is more than just managing daily operations; leaders shape and enact
policies, make data-based decisions, monitor achievement, implement curricula, and oversee faculty
development. Microsoft Copilot helps education leaders accomplish many of these time-consuming tasks.

Microsoft Copilot Copilot for Microsoft 365 Copilot in Windows

Simplify Windows 11 device

Increase productivity when Use Microsoft 365 apps and management and perform

Plan
completing administrative files to complete specialized activities like:
duties like: tasks like:
• Troubleshoot audio/video
• Research and compare • Summarize internal issues during lessons
curricula state reports
• Adjust device notifications
• Outline an agenda for • Auto-draft messages when planning
professional learning to faculty
• Find documents without

Implement
• Summarize online articles • Create visualizations knowing the file name
or PDFs from spreadsheets

Copilot prompt

AI Navigators
Summarize the 2024 National Educational Technology Plan and Include sections on the digital
use divide, digital design divide, and the digital access divide. Provide a 1-2 sentence definition
of each digital divide and list 5 steps to take to address the divide in each section. The summary
should be written in plain language that’s understandable by educators. Cite any source material.
Research

Navigate back to
14 Microsoft Education AI Toolkit
table of contents
Copilot for educators

Overview
As the people most directly responsible for student learning, educators spend the bulk of their working
hours writing lesson plans, assessing understanding, facilitating classroom activities, and completing
administrative duties. Microsoft Copilot makes common educator tasks more manageable and efficient.

•
Microsoft Copilot Copilot for Microsoft 365 GitHub Copilot

Increase productivity and Use Microsoft 365 apps Deploy an AI-powered

save time completing duties and files to accomplish coding assistant that

Plan
like: specialized tasks like: supports computer science
instruction:
• Create a course syllabus • Recap Teams meetings for
absent students • Provide students with
• Write a lesson plan that just-in-time coding support
differentiates instruction • Auto-draft emails for
families • Debug complicated
• Level text for emergent programs and refactor
readers • Create a rubric from a code
lesson document

Implement
• Help students document
change logs

Copilot prompt

AI Navigators
You are an AI with expertise in physics. Your task is to provide five diverse analogies that can help
explain Bernoulli’s Principle to high school students preparing for their state exams. The analogies
should be simple, concise, and cater to a range of student interests and experiences. Remember,
your goal is to aid their understanding of the principle, not to introduce more complexity.
Research

Navigate back to
15 Microsoft Education AI Toolkit
table of contents
AI for students

Overview
By equipping students with the knowledge and tools needed to safely interact with AI products in
the classroom, we can better prepare them for the real-world challenges they will encounter. Recent
research, sponsored by Microsoft and conducted by the International Data Corporation (IDC), reveals
significant insights into the widespread adoption of AI in organizational settings. The study indicates
that 71% of organizations have already integrated AI into their operations, with a staggering 92%
completing their AI deployments within a year or less. Additionally, for every dollar invested in AI,
companies are witnessing an average return of $3.5. However, despite the evident benefits, over
half of these organizations (52%) cite a shortage of skilled workers as a major obstacle to effectively
implementing and expanding AI initiatives1. This underscores the pressing need for educational
initiatives that help address this gap and empower students to gain requisite expertise to thrive in the
digital economy.

Plan
Microsoft's suite of AI-powered solutions plays a pivotal role in fostering inclusive and equitable
learning environments. Students gain access to tools such as Copilot and Learning Accelerators at no
extra cost, provided they utilize a school-issued Microsoft account and meet the necessary age criteria.
This commitment to accessibility ensures that all students, regardless of background or financial means,
can leverage cutting-edge technology to enhance their educational journey.

Students Students All

Implement
18+ 13+ Students

Microsoft Copilot
•* •**
Copilot for M365
•
GitHub Copilot
• •
Learning
Reading Coach
• • •
Accelerators
Search Coach
• • •

AI Navigators
**Institutions must provision accounts through Microsoft Entra ID to support student access.
*Microsoft is expanding Microsoft Copilot access to younger learners through a private preview in the coming months.

Most of Microsoft’s Learning Accelerators provide AI-powered learning support for students. Two
notable examples are Reading Coach and Search Coach. These tools, combined with direct instruction
and the guidance of educational practitioners, assist primary and secondary students in developing
critical reading, writing, listening, and speaking skills necessary for using AI effectively.
Research

1
Toward Skilling Excellence: The IDC 2022 Global IT Skills Survey

Navigate back to
16 Microsoft Education AI Toolkit
table of contents
Learning Accelerators

Overview
Get to know Microsoft Learning Accelerators

Plan
Reading Progress Search Progress Speaker Progress
Tracks student reading skills Enables educators to guide Provides data-driven
and provides educators students' information insights on students'
with actionable insights for literacy skills by monitoring speaking skills.
targeted improvement areas. their search activity and
query quality resources.

Reading Coach Search Coach Speaker Coach

Implement
Offers AI-powered,
Fosters information literacy Offers real-time feedback
personalized reading fluency
by coaching students to on public speaking
practice, enabling learners
develop effective search elements within PowerPoint
to co-create stories and
queries and identify and Teams.
practice challenging words.
reliable resources.

AI Navigators
Reflect Math Progress*
Aids educators in creating practice questions
Reflect encourages students to identify and
and analyzing students’ challenges, facilitating
express emotions and provides educators
personalized feedback and support.
with insights to offer support.
Math Coach**
Enhances math learning with real-time feedback
and personalized practice for students.
Research

Education Insights
Integrates data across Learning Accelerators to equip educators with a
comprehensive view of each student's academic journey.

*Entering private preview

**Coming soon
 Overview
Plan
Implement
Introduction to AI

Section 2
Plan

AI Navigators
Research

Navigate back to
table of contents
Leveraging responsible AI frameworks

Overview
The increasing use and application of generative 1. Why might you want to use generative
AI in education holds great promise. Ensuring AI tools?
the responsible use of generative AI in education
rests on a combination of the policies, guidelines, 2. How does your institution currently
and frameworks you set and the tools you choose handle the adoption of new technologies?
to adopt; in essence, our shared responsibility. Will that existing model work for AI?
For example, safeguarding student data is
dependent on both good policies and effective 3. Does it make sense for your institution to
training so educators and students know what to create one AI policy or modify and adapt
do and how to do it. For leaders, it is imperative many different ones?

Plan
to collaborate with educators, policymakers, and
stakeholders to empower all users with the best 4. What systems do you have in place to
that AI technology has to offer while upholding ensure equitable application of these
the highest standards of responsibility. policies as you consider using AI tools?

This section of the AI Toolkit highlights the 5. What are the legal considerations that you
importance of responsible AI practices, provides need to address?
insight into Microsoft’s approach to responsible
Ai and suggests ways ways to help you get your There are a variety of AI frameworks that might
institution and its policies AI-ready. Questions help you answer these questions. For example,

Implement
you may want to consider as you review these the resources from Teach AI include sample
frameworks and policy examples include: school guidance materials and frameworks
including a focus on reviewing AI output,
checking for incorrect information, transparency,
safety, and respect.

AI Navigators
Copilot prompt

Assume the role of an education institution leader such as a provost or superintendent for
a medium-sized institution. Provide a list of six policies, frameworks, or guidelines (such as
Acceptable Use Policies) that should be reviewed and considered for revision to allow for the use of
generative AI responsibly and ethically. Additionally, describe three different types of AI use policies
that could be developed by schools, universities, or ministries of higher education for reference.
Research

1
teachai.org/

Navigate back to
19 Microsoft Education AI Toolkit
table of contents
Principles of responsible AI

Overview
It is important to consider responsible AI principles when implementing AI in education to ensure
these technologies are used responsibly, safely, and in ways that enhance educational opportunities
for students while preparing them for the future. Responsible AI principles can be used to help inform
policies and usage guidelines adopted by districts, states, and ministries of education. They can help
ensure that AI is used in ways that are fair, transparent, and respect the privacy of students and staff. At
the core of Microsoft’s AI work are six key principles that define and guide responsible AI use.

Plan
Microsoft’s Responsible AI Principles

Fairness Reliability Privacy Inclusiveness

Implement
& Safety & Security

Transparency

Accountability

AI Navigators
Research

Navigate back to
20 Microsoft Education AI Toolkit
table of contents
Practical steps for education leaders

Overview
Translating these guidelines and frameworks into actionable steps is one of the central challenges that
education and leaders face when considering generative AI. As you navigate the evolving landscape of
generative AI, there are some easy places to begin so that you can build trust within your organization
that your AI roll-out will be fair and responsible.

Suggested actions Resources

Plan
Review policy documentation
Rethinking Acceptable Use
Revise policies to address like Acceptable Use Policies to
Policies in the Age of AI,
generative AI incorporate language about
District Administration
the use of generative AI.

Set guidelines for the

Incorporate AI into teaching responsible use of AI tools in Integrating Generative AI into
and learning lesson planning and course Higher Education, Educause
creation.

Implement
Establish monitoring and
Create a plan to monitor and ChatGPT and Beyond,
evaluation standards for AI
assess the use of AI. Common Sense Education
content

AI Navigators
Copilot prompt

As a leader in a medium-sized educational institution, such as a provost or superintendent, you

are tasked with preparing your institution for the implementation of generative AI. Draft a ten-
step plan for integrating generative AI in your educational institution. Focus on policy updates,
implementation strategies, and evaluation methods to ensure a smooth transition.
Research

Navigate back to
21 Microsoft Education AI Toolkit
table of contents
Institutional policy considerations

Overview
Practical preparation
Establishing policies creates structure and guidelines for your faculty, staff, students, and community.
Before you get started, consider these practical suggestions:

Start now: Your students and staff are likely using AI already and

Plan
need guidance. Create initial policies and iterate as you go.

Identify key areas of need and critical questions that will guide
your process.

Implement
Establish what needs policy and what doesn’t. Focus on the
largest areas of impact.

Learn from peers and familiarize yourself with resources like the
TeachAI toolkit, developed with support from Microsoft.1

AI Navigators
As your school or institution develops its AI strategy, it's natural to shift your focus to areas that are
affected, especially policies that may need updating to address recent changes. Start by consulting state
or national guidelines and requirements and reviewing your existing policies. Then, consider curating a
set of exemplary policies that can be customized to meet your specific needs.
Research

1
teachai.org/toolkit

Navigate back to
22 Microsoft Education AI Toolkit
table of contents
Organization policy considerations

Overview
Crafting, updating, and approving new policies is a critical task. We encourage you to take an iterative
approach to policy making that emphasizes ongoing improvement. In the case of AI, schools are still
exploring the full extent of its use and impact, and a successful policy is one that is regularly reviewed
and revised to meet the current needs of the school and community.

This section begins by examining a hypothetical academic integrity policy and outlines how a
school could enhance it for better guidance on students’ use of AI. The latter part offers concise
recommendations for reviewing other policies.

Academic integrity policy

Plan
Early research reveals that students who have access to explanations created by generative AI perform
better than students who only have access to correct answers.2 AI’s impact goes beyond tool usage
and user considerations. It involves how every member of the school or higher educational community
adopts and uses AI tools. Your school, like many others, aims to define clear guidelines for academic
integrity considering increased student use of generative AI tools. It’s crucial to evaluate the effects of
policies, whether highly restrictive, fully encouraging, or a mix of both, on students, their education, and
the overall fulfillment of academic standards.

Implement
Questions to lead your discussion Policy spotlight

• Are your students allowed to use AI South Australia’s Department for Education
on assignments? led a pilot program that introduced a custom
chatbot for students to use. Their policy provides
• Which policy model will guide structure and guidance around how learners can
your practice?3 responsively use generated content.4

AI Navigators
• What impact will that have on your current
academic integrity policy?

Research

2
papers.ssrn.com/sol3/papers.cfm?abstract_id=4641653
3
oue.fas.harvard.edu/ai-guidance#:~:text=POLICIES%20FOR%20THE%20USE%20OF%20AI%20IN%20COURSES
4
education.sa.gov.au/parents-and-families/curriculum-and-learning/ai#:~:text=AI%20and%20AI%2Denabled%20technology,security%20and%20students’%20
learning%20needs.

Navigate back to
23 Microsoft Education AI Toolkit
table of contents
Academic integrity highlight: Refining a policy

Overview
Below is one possible evolution of an academic integrity policy. Each step in the evolution includes a
sample policy which is followed by a quick analysis of its effectiveness.

Initial policy Revised policy

Presenting another person’s work as your own is Presenting another person’s work or content
an act of dishonesty. This behavior undermines created by a generative AI tool as your own is
your integrity and contradicts the principles an act of dishonesty. This behavior undermines
upheld by [our institution]. We maintain the your integrity and contradicts the principles
belief that academic success is contingent upon upheld by [our school]. We maintain the belief
the dedication you invest in your studies. that academic success is contingent upon the

Plan
dedication you invest in your studies. We expect
Analysis you will approach your assignments honestly, as
your work reflects your capabilities.
This policy addresses human-authored texts.
Given the many ways that students can Analysis
use generative AI tools, clear guidance on
responsible AI use is essential to maintain This policy covers generative AI. It broadens the
academic integrity and prevent plagiarism. range of permitted uses for students beyond
mere assignment copying but does not offer

Implement
appropriate uses for AI. We recommend setting
guidelines for additional generative AI uses like
revising work, seeking formative feedback, and
utilizing AI as a brainstorming partner.

Leadership teams can create prompts to assess existing policies for improvement and explore various
wording options. For instance, Copilot can analyze a revised policy, review it for potential biases, and
request a simplified version in plain language accessible to all student and community groups.

AI Navigators
Apply your learning

Open your institution's academic integrity policy in the Edge browser. Open Copilot sidebar from
the top right and enter the prompt below:

As the CAO of a school district, analyze our existing academic integrity policy, focusing on AI's ethical
use by students. Evaluate the policy's current consideration of implicit biases, cultural, and socio-
economic diversity. Suggest concrete, actionable improvements to enhance inclusivity, fairness, and
clarity, ensuring the policy is understandable and accessible to all students. Provide examples of best
practices from other policies and include a revised policy draft incorporating these elements.
Research

Navigate back to
24 Microsoft Education AI Toolkit
table of contents
Sample policy developed with Copilot

Overview
At [our school], we prioritize academic integrity as a core principle. We anticipate that all students will
complete their assignments with honesty, understanding that their work showcases their abilities.
Additionally, we emphasize responsible AI usage, which involves thoroughly reviewing content rather
than merely copying and pasting generated material. To promote responsible AI use and ensure
fairness, we provide the following guidelines:

Attribution and AI content Implicit bias awareness

• When using AI-generated content, always • Our academic integrity process aims to be
provide proper attribution. unbiased and consider individual circumstances.

Plan
• Presenting AI-generated work as your own is Education and resources
strictly prohibited. Faculty will set clear • We offer workshops, tutorials, and online
expectations regarding responsible AI use for resources on citation practices and
their class using approved categories. responsible AI use.
Sample policy language:
✓ Highly restrictive: No AI use is allowed. • Students are encouraged to learn and apply
✓ Fully encouraging: AI is fully available these principles.
for student academic use. No restrictions.
✓ Hybrid: AI use is for brainstorming and Equitable enforcement

Implement
AI tutorials, but not for submitted • Violations will be addressed consistently,
assignments. regardless of socio-economic status or
cultural background.
For additional informtion or language, refer
to the AI guidance & FAQs from Harvard.5 • Fairness and equity guide our approach.

Cultural sensitivity • If assignments allow or require AI use or

• We respect diverse cultural norms related to specific tools, those tools must be readily
collaboration and attribution. available and provided for each student.

AI Navigators
• Students should be aware of these differences By adhering to these guidelines, students
and adapt their practices accordingly. contribute to a respectful and inclusive academic
community. Let us work together to foster
integrity, learning, and growth.

Analysis

This policy outlines student expectations, emphasizing the many ways that students may use generative
AI tools in their workflow. It reflects the school's dedication to fairness and outlines efforts to teach
students responsible AI skills.
Research

5
oue.fas.harvard.edu/ai-guidance#:~:text=POLICIES%20FOR%20THE%20USE%20OF%20AI%20IN%20COURSES

Navigate back to
25 Microsoft Education AI Toolkit
table of contents
Additional policy considerations

Overview
After exploring the process of evaluating and drafting an academic integrity policy, let's apply the same
strategies to other policy areas. Consider how your team can implement the methods discussed earlier
in these new contexts.

Data protection and privacy Staff and faculty use

Large language models (LLMs) depend on user AI tools can enhance educator efficiency and
data to produce results. It’s important that personalize student content. Clear guidelines
data and privacy are core considerations when for AI tool usage are highly recommended. The
approving AI tools. Further, take the time to generative AI benefits for educators range from

Plan
transparently communicate to students, staff, and creating lesson plans, content curation, task
the community about how the school and tools automation, and generating communication (e.g.,
use and protect their data. emails). Use these guiding questions to lead
your discussion:
Use these guiding questions to lead your discussion:
• How might we improve learning through
• What does student privacy mean in the AI era? using AI for instructional purposes?

• How well do our data protection and privacy • What instructional uses do we want to
policies align with legal regulations. encourage? What might we restrict?

Implement
• How do we communicate our data usage • How will we support our staff
policies and practices to students, staff, and throughprofessionallearning?
families? Is there an opt-out option?

Policy spotlight

AI Navigators
Educators in Wichita Public Schools have used Copilot to develop instructional resources and
individualize their students’ learning. In this scenario, the educators would benefit from a combination
of focused professional learning alongside established guidelines in the district’s policies.
Research

Navigate back to
26 Microsoft Education AI Toolkit
table of contents
Classroom syllabi Accessibility and Universal Design

Overview
for Learning (UDL)
Educators utilize syllabi to communicate
expectations, instructional resources, and
AI tools have the potential to make learning
experiences, often incorporating academic
more accessible for all learners. From simply
integrity statements. Consider offering educators
adapting content into more accessible formats
a standardized statement for inclusion or
to building entirely new instructional content
adaptation in their syllabi. Use these guiding
based on an individual student’s needs, AI tools
questions to lead your discussion:
offer great promise for making learning more
accessible and equitable. Use these guiding
• What message needs to be included on all
questions to lead your discussion:
class syllabi?

Plan
• What are the accessibility needs of our students
• How can this statement reiterate broader policies?
and staff?

• To what extent can educators adapt the
• How might AI tools help us increase our ability
statement for their classes?
to make learning accessible for all?

• What government guidelines need to be

followed as we evaluate AI tools and
structure our school’s AI program?

Implement
Policy spotlight

California State University San Marcos expanded the traditional use of “accessibility” by creating AI tools
that enabled custom communication to better reach diverse groups of students such as first-generation
college learners.

AI Navigators
Research

Navigate back to
27 Microsoft Education AI Toolkit
table of contents
 The importance

Overview
of a secure foundation
Education leaders know that protecting data and preventing
“Artificial Intelligence
cyberattacks are essential for safe, secure, and effective learning
must be safe and environments. Data-rich organizations like schools, universities,
secure. Meeting this and ministries of education are targeted by bad actors and
cybercriminals, as shown by increased attacks and new social
goal requires robust,
engineering threats.
reliable, repeatable,
and standardized • Approximately 80% of all malware exploits occur in the field

Plan
of education, a figure that is substantially higher than any
evaluations of AI
other industry.
systems, as well as
policies, institutions, • Microsoft blocked an average of 4,000 password attacks
per second in 2023.
and, as appropriate,
other mechanisms to • The global average cost of a data breach was $4.45 million
test, understand, and in 2023.2
mitigate risks from
The U.S. Cybersecurity and Infrastructure Security Agency (CISA)
these systems before

Implement
launched a campaign to address the spike in cyberthreats
they are put to use.” impacting education over recent years. Beginning with the
Protecting Our Future report, federal security experts and the
Department of Education outlined key recommendations that
—U.S. Presidential increase security and privacy in schools.3 In late 2023, President
Biden codified additional ways AI systems must be secured
Executive Order 141101
when he signed the Executive Order on the Safe, Secure, and
Trustworthy Development and Use of Artificial Intelligence.4
The Executive Order called for public and private organizations
to make sure that AI systems are resilient against misuse and

AI Navigators
modifications, function as intended, ethically developed and
deployed in a secure manner.

Schools and universities across the country are listening to

the government’s calls for increased cybersecurity protection
and close examination of security and privacy in AI systems.
More and more states are adopting policies for safely using AI
systems in K–12 school districts with help from companies like
Microsoft and government agencies. Microsoft is also working
closely with higher education institutions like the University of
Michigan to deploy copilots that are as secure as other existing
Research

1
whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence/
2
ibm.com/reports/data-breach
3
cisa.gov/resources-tools/resources/report-partnering-safeguard-k-12-organizations-cybersecurity-threats
4
whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence/
5
er.educause.edu/articles/2024/2/how-and-why-the-university-of-michigan-built-its-own-closed-generative-ai-tools

Navigate back to
28 Microsoft Education AI Toolkit
table of contents
infrastructure.5 And with the launch of the Secure Future Initiative (SFI) in 2023, Microsoft accelerated
work on their AI-based cyber defense systems so that organizations like schools and institutions have

Overview
machine-speed security.

But what can you do if you want to provide secure AI systems but don’t know where to start?

This section of the AI Toolkit provides you with steps to follow to help you safely and securely
implement generative AI tools. You will also learn how Microsoft’s AI systems and A3/A5 Microsoft
365 Education plans provide security tools that give you control and protection in managing AI in your
school's infrastructure.

Step 1 Step 2 Step 3 Step 4 Step 5

Plan
Understand Identify Establish data Determine Develop an
the importance outcomes and governance, data privacy incident
of a respsonsible data sources for roles, and procedures and response plan
AI framework AI systems responsibilities safeguards

Implement
Understand the importance of a responsible
Step 1
AI framework

AI Navigators
The first step towards a secure AI foundation begins with reviewing Responsible AI practices and
understanding any potential issues with using AI in your educational setting. Keep the following
questions in mind as you evaluate AI systems:

1. Does the AI system follow responsible AI practices?

2. How did developers and designers follow responsible AI practices when developing their solutions?

3. Does the AI system align with jurisdictional laws and regulations for my school or institution?
Research

Navigate back to
29 Microsoft Education AI Toolkit
table of contents
Responsible AI implications

Overview
It’s important to research and understand how generative AI systems are built before developing or
deploying a tool in your environment. Consider an example from K–12 education:

Stanford researchers found that seven different GPT systems falsely identified over 50% of
Chinese students’ essays for an 8th grade Test of English as a Foreign Language (TOEFL) as “AI-
generated” when they were not.6 There were no inaccuracies when the GPT systems analyzed
8th grade essays written by native English speakers.

Plan
Fairness issues–like assessing writing differently when a learner is a non-native speaker–is a Responsible AI
principle that should be examined when considering AI systems. Vendors should be able to articulate how
their systems address all of the Responsible AI principles, so consider asking them questions about the
development process or the steps that are taken to ensure the AI system remains responsibly designed.

Implement
Microsoft develops AI systems using Responsible AI principles. From development to deployment and long
after, engineers and technicians follow the Microsoft Responsible AI Standard to check that AI systems
provide “similar quality of service for identified demographic groups, including marginalized groups.

Before moving on, evaluate how AI systems generate content and conduct research on reports of bias
in outputs.

AI Navigators
Research

6
doi.org/10.1016/j.platter.2023.100779

Navigate back to
30 Microsoft Education AI Toolkit
table of contents
 Identify outcomes and data sources for

Overview
Step 2
AI systems

Next, it can be helpful to create a list of the desired outcomes you want the AI system to accomplish
and what data might be required. Keep the following questions in mind:

1. What are some pain points in your school or institution?

2. What needs do community members have that an AI system might be able to address?

Plan
3. How does the AI system use data to generate responses?

Goals
AI systems perform different functions and have different capabilities. For example, Microsoft
Copilot helps people increase their productivity and creativity whereas Copilot for Security provides
critical incident response, threat hunting, intelligence gathering, and posture management for IT

Implement
administrators. Knowing what you want the AI system to accomplish–your goals–will help you find the
right solutions for your institution.

It’s helpful to begin by making a hierarchical list of needs or pain points that an AI system might
address. Place the most urgent items at the top of the list. You might want to ask colleagues from other
departments to offer input on what you identified.

After you have your list, consider re-writing the pain points into goal statements. For example:

Pain point: IT administrators struggle to prioritize threats because of the number of signals
that emerge each day.

AI Navigators
Goal statement: Any security-focused AI system should help administrators prioritize threats
and give guidance on steps to take to respond appropriately.

Compare your goal statements against the capabilities of any AI systems that you are considering.
If there’s alignment between your goal and the AI system, review how it addresses some of the
responsible AI requirements that you learned about in Step 1.
Research

Navigate back to
31 Microsoft Education AI Toolkit
table of contents
Data sources

Overview
AI systems use data to generate responses. Sometimes AI needs access to files on your computer or
network, like when you ask Copilot for Microsoft 365 to summarize important points from a notes
document. There are also instances when AI requires access to critical systems to perform specialized
tasks. For example when you build a custom Copilot with Azure AI Studio, you have options to connect
private data sources to increase customization and deliver fine-tuned responses. Regardless of which
AI system meets your goals, be mindful of the data requirements and laws that govern data privacy for
your jurisdiction.

The following activities can help you make informed decisions about data handling:

• Form a committee that includes compliance officers, security administrators, and other leaders.

Plan
Review and revise the goals you wrote, identify any data sources required in the AI systems under
consideration, and list any compliance requirements that must be met before implementation.

• Draft a list of data governance questions that you can ask vendors. Key topics include what data
sources are required, how data is kept safe, and what mechanisms help you manage risk.

• Microsoft provides commercial data protection for Microsoft Copilot and Copilot in Windows. Both
copilots do not save prompts or responses, data is not used to train models, and no one at Microsoft
can view school data when a user logs in with a Microsoft-issued account.

Implement
Before moving on:

✓ Align your stated goals with the capabilities of AI systems under consideration.

✓ Determine what data is required to use an AI system.

✓ Evaluate data requirements against any laws or regulations that govern your school or institution.

AI Navigators
Research

Navigate back to
32 Microsoft Education AI Toolkit
table of contents
 Establish data governance, roles,

Overview
Step 3
and responsibilities

After you have identified an AI system that’s secure, compliant, and addresses a goal, begin exploring
data governance for your school or institution. Keep the following questions in mind:

1. Does my school or institution have the infrastructure required for AI applications

to access data securely, quickly, and at scale?

2. What infrastructure and resources are available to support AI deployment?

Plan
3. Who is going to be responsible for ongoing monitoring, troubleshooting,
and communication?

Develop a plan and consider roles

It’s important to assess your infrastructure’s readiness for safe, secure AI experiences. Sometimes
you will find an AI system that meets or exceeds the security requirements dictated by laws, but your
infrastructure or staff are not prepared to safeguard and monitor interactions. Having a plan will help
you think through data governance issues that might be specific to your school or institution.

Implement
Some questions to consider as you develop your plan:

• Would it be better to buy a pre-built AI system, develop AI applications in-house, or update

existing AI systems?

• Should data for AI systems be stored on-premises or in the cloud?

• What data platform architecture is necessary to run the AI system? Do the requirements match
what is allowed by law for my school or institution?

AI Navigators
In addition to assessing infrastructure capabilities, you also need to examine the extent to which your
IT administrators can monitor and assess any AI system that goes into your environment. The Cyber
Security and Infrastructure Security Agency (CISA) recommends establishing an incident manager,
technology manager, and a communication manager to oversee AI systems.7
Research

7
cisa.gov/sites/default/files/publications/Incident-Response-Plan-Basics_508c.pdf

Navigate back to
33 Microsoft Education AI Toolkit
table of contents
 Overview
The incident manager The tech manager is the The communication
leads the response when an subject matter expert for manager communicates
incident with an AI system all AI systems. They should with internal and external
occurs. They manage have knowledge of AI, data stakeholders about important
communication flows and security, and appropriate decisions or incidents.
delegate tasks but do not response measures.
perform any technical duties.

Plan
If you can’t identify who might serve in these capacities, consider hiring someone who might have the
expertise to perform the functions. These roles are important! Assigning people to roles and including
them in any implementation plans will help you gather diverse perspectives and build teamwork well
before the school community begins using the AI system.

The A3/A5 Microsoft 365 Education plans and security add-ons include applications that help monitor
AI activities and data flow. For example:

• Microsoft Defender for Cloud: Monitor AI system usage across cloud, multi-cloud, or hybrid

Implement
infrastructures, understand associated risks, and approve or block access by browsing a catalog of
400+ generative AI applications.

• Microsoft Purview: Detect data security risks in Copilot for Microsoft 365 through Purview’s AI hub.
The AI hub aggregates usage statistics and applies a risk level to over 100 of the most common AI
applications. Purview also uses sensitivity label citation and inheritance for additional security with
AI systems.

• Microsoft Purview eDiscovery: Identify, preserve, and collect relevant AI data and interactions
litigation, investigations, audits, and inquiries.

AI Navigators
Before moving on:

✓ Evaluate your organization’s Zero Trust security posture and overall capabilities.

✓ Determine data management requirements like cloud-based or on-premises storage.

✓ Identify colleagues who have the expertise to serve in AI-specific roles.

Research

Navigate back to
34 Microsoft Education AI Toolkit
table of contents
 Determine data privacy procedures

Overview
Step 4
and safeguards

Data privacy is a central concern when using AI systems or any type of technology in K-20 settings.
Understanding how data is kept safe and secure is one piece of information that you need to assess. It’s
also important to review internal policies and identity access protocols prior to deploying an AI system.
Keep the following questions in mind:

1. What are the known privacy risks with the AI system?

Plan
2. How is data shared, used, and stored in the AI system?

3. How do people access and use the AI system?

Privacy impact assessments

Privacy impact assessments (PIA) are tools that government agencies use to evaluate information
technology (IT) systems for privacy risks and options for mitigating these risks. Learning how agencies
approach evaluating privacy and data risks in a privacy impact assessment can help you assess solutions

Implement
that you are considering. A privacy impact assessment typically includes:

• Known privacy risks;

• Options for mitigating known privacy risks;

• Instructions on how to properly handle privacy issues;

• Processes for analyzing the legal compliance with privacy laws and regulations;

• Documentation on the flow of personal information; and

AI Navigators
• Public assurances that personal information is protected.

If you can address each one of these points when you evaluate an AI system you will have enough
information to make an knowledgable decision about data privacy protection.
Research

Navigate back to
35 Microsoft Education AI Toolkit
table of contents
Privacy policies

Overview
Vendors should clearly articulate how data is used, stored, and shared when discussing AI solutions with
you. After all, schools and institutions are charged with helping protect the privacy of students, so AI
solutions need to provide adequate security and protection built into the application design.

It’s wise to ask vendors for data privacy documentation or a legal Privacy Policy for any AI
systems you are considering. Reviewing data privacy and security statements can help you
make informed, legal decisions about what’s best for your school or institution. For example,
Microsoft publishes how data is used in each one of its AI systems:

Plan
Microsoft Copilot Copilot in Windows Copilot for Microsoft 365

Implement
Copilot for Security Azure AI Studio

Infrastructure settings
Your computing infrastructure and security services can add additional layers of privacy protection
depending on how everything is configured. Offering secure identity access protocols and creating user
polices are two ways to increase privacy without relying on built-in protection within AI systems. Take
time to speak with IT administrators so that you are aware of your system’s capabilities.

AI Navigators
Microsoft offers two solutions that help you maintain data privacy when using AI systems:

• Microsoft Entra ID: Manage access to Microsoft Copilot and underlying data with secure
authentication procedures and risk-based adaptive policies.

• Intune for Education: Apply security, configuration, and compliance policies to devices so that
school-issued endpoints have baseline protection when working with AI systems.

Consider asking vendors how their AI systems integrate with Microsoft solutions or whatever products
you use to control access and manage policies.
Research

Navigate back to
36 Microsoft Education AI Toolkit
table of contents
Before moving on:

Overview
• Use the parts of a privacy impact assessment to assess your institution's data privacy protocols.

• Obtain a copy of any privacy and security documentation for the AI system.

• Evaluate your existing infrastructure for ways to bolster identity access control and implement
security policies.

Develop an incident response plan to

Step 5
address issues that arise

Plan
Having an incident response plan ensures that you can respond accordingly when an issue arises.
Incidents occur in even the most secure infrastructure, so having a response plan before you launch
an AI system will help you think through necessary logistics and procedures ahead of time. Keep the
following questions in mind:

1. What constitutes an incident with an AI system?

2. What parts go into an incident response plan?

Implement
3. Who should be notified when an incident occurs?

Incident response plans

Before diving into incident response plans, you should understand what constitutes an incident. Microsoft
defines an incident as a group of correlated alerts that humans or automation tools deem to be a genuine
threat. Although one alert on their own might not constitute a major threat, the combination of alerts
might indicate a possible breach.

Well-designed, secure AI systems that run inside carefully managed infrastructure still face threats that

AI Navigators
lead to incidents. Some common points of failure include:

✓ Security breaches that give cyber criminals access to sensitive data;

✓ Unintentionally disclosing private information that would not otherwise be shared; and

✓ Discriminatory or misleading information in responses.

Research

Navigate back to
37 Microsoft Education AI Toolkit
table of contents
Developing an incident response plan that handles issues like these will enable you effectively address
the issues that will arise. CISA recommends schools and institutions layout a 6-stage incident response plan.8

Overview
1. Preparation 2. Detection and analysis 3. Containment
Document and share Define processes that are Develop an approach to
policies and procedures for implementable, gather contain or minimize threats
incident response, configure baseline information and identify known security
security systems to detect to monitor and detect containment strategies for
suspicious and malicious anomalous or suspicious common incidents.
activity, create roles and activity, and outline the
assign responsibilities, and differences between an
educate users about the AI authorized use and

Plan
system. an incident.

4. Eradication and recovery 5. Post-incident activity 6. Coordination

Outline how artifacts Outline how incidents Identify who needs to be
from an incident can be should be documented informed when an incident
eliminated, create mitigation and reported, identify steps occurs depending on the
steps to address exploited for hardening security, and

Implement
severity of the threat.
vulnerabilities, document develop ways to share and
how evidence will be apply lessons learned.
collected, establish a regular
back-up plan, and list steps
for recovery and return to
normal operation.

A great way to create an incident response plan is to form a committee with colleagues who have
expertise in each area. If you identified an incident manager, technology manager, or communication
manager in step 3, invite these people to contribute to the incident response plan for AI systems. For

AI Navigators
additional information on developing your own cybersecurity incident response plan, check out CISA’s
Incident Response Plan (IRP) Basics or the K12 Six Essential Cyber Incident Response Runbook.9,10

Research

8
2.ed.gov/notices/pia/index.html
9
cisa.gov/sites/default/files/publications/Incident-Response-Plan-Basics_508c.pdf
10
static1.squarespace.com/static/5e441b46adfb340b05008fe7/t/62cc8e3843251c6d2b2cb0a5/1657572921164/K12SIX-IncidentResponseRunbook.pdfb

Navigate back to
38 Microsoft Education AI Toolkit
table of contents
Data governance

Overview
Ensuring the security and integrity of your effective, and innovative educational environment.
data assets has become a top priority for This includes responsibly adding AI into our
organizations across all sectors. Creating a schools, ensuring it not only improves learning
strong security posture includes a well-defined but also respects and safeguards the privacy and
data governance framework. Data governance integrity of all data and individuals involved.
and security are fundamentally intertwined,
each reinforcing the other to safeguard the Generative AI has many potential applications
confidentiality, availability, and integrity of data. in education, however, generative AI also poses
By combining effective data governance with significant challenges for data governance and
robust security measures, organizations can privacy, as it may involve processing sensitive or

Plan
defend against a wide range of cyber threats, personal data. This document provides a practical
ensuring that their data is both well-managed guide for you on addressing these challenges and
and highly secure. implementing generative AI responsibly.

Cloud data consolidation Data governance in AI

In an educational setting, migrating data to cloud Data governance is the process of defining and
storage offers significant advantages. A cloud implementing policies, standards, and practices
data governance framework acts as a strategic for managing data quality, security, and

Implement
blueprint, guiding the storage and management compliance. This also helps establish processes
of your data in the cloud. Tools enable your team that keep your data secured, private, accurate,
to monitor and understand data movements, and usable throughout its life cycle.
while your governance framework establishes the
rules, roles, procedures, and processes for securely A healthy data governance strategy is crucial
managing and controlling these data flows within for any organization that relies on data to
cloud storage. This approach helps ensure safe improve decision-making and ensure successful
and efficient data handling in schools. outcomes. When collecting vast amounts of data,
establish a strategy that manages risks, reduces
costs, and executes your objectives effectively.
Data governance and

AI Navigators
Data governance is essential for ensuring that

privacy needs generative AI is used for legitimate and beneficial

purposes, and that the risks of data misuse,
leakage, or corruption are minimized.
We understand that, as Information and
Instructional Technology leaders, you face a
delicate balance between needing to manage data
safety and privacy while adopting new technology.
Your expertise is pivotal in creating a secure,
Research

Navigate back to
39 Microsoft Education AI Toolkit
table of contents
In the context of AI, data governance transcends • Data management: Reflect on your data
regulatory compliance, serves as the foundation management practices, specifically regarding

Overview
for responsible and effective AI deployment in the handling, storing, and retrieving of data.
education. Evaluate your existing framework with How have you ensured that your data
these key questions in mind: architecture not only meets the operational
needs of AI systems but also aligns with the
• Data Security: Regarding your current strategic goals of the educational institution
security measures for your organization's
data, can you detail the process of Along with managing these key components,
documenting these security measures and you often:
how you manage data loss prevention efforts
within your organization • Craft and enforce data governance frameworks
that align with your organization’s mission

Plan
• Data policies: Have you considered the and goals.
development and implementation of
comprehensive data policies to govern • Set up teams or committees responsible for
the use and sharing of data within your AI various aspects of data governance.
systems? How do these policies ensure
transparency and uphold ethical standards, • Evolve your organizational policies as AI
especially when handling sensitive advances and changes.
student data
• Stay on top of legal requirements like FERPA in
• Data sources: Have you reflected on the the United States and GDPR in Europe.

Implement
diversity and reliability of the data sources
underpinning your data governance It requires a blend of technical expertise,
framework? How do these sources impact the foresight, and leadership to ensure your
integrity and effectiveness of your data organization remains compliant and leverages its
management strategies data effectively.

• Data compliance: Consider how data

compliance, beyond mere adherence to
regulations, plays a foundational role in your
data governance framework. How might this
Copilot prompt

AI Navigators
influence the efficacy and responsibility of
your data management practices
Examine the overlooked elements in
implementing data governance in public
• Data quality: Think about the measures
educational institutions. Address the roles of
you've taken to ensure the accuracy,
stakeholders like IT leaders, administrators
consistency, and reliability of data,
and educators, and the challenges with
particularly in an educational context. How do
types of data such as student performance
you verify the integrity of student
and privacy. Provide actionable insights for
performance data, demographic information,
improved governance.
and the relevance of learning materials
Research

Navigate back to
40 Microsoft Education AI Toolkit
table of contents
Data privacy considerations in AI-driven education

Overview
Safeguarding students’ personal data is paramount, especially as many AI applications, particularly those
that are data-driven or utilize machine learning, require access to large datasets to learn, adapt, and
provide personalized experiences. The challenge is to balance the functionality and benefits of these
AI applications with the imperative to protect sensitive data. Data privacy helps ensure individuals have
control over their personal data's collection, usage, and sharing. Key considerations include:

Student, educator, and faculty data privacy

The integration of AI in education necessitates careful management of both student and faculty data,
encompassing academic performance, learning behaviors, and sensitive personal details. Safeguarding

Plan
the privacy of this data is crucial to preventing breaches that could jeopardize the privacy of students
and educators alike, as well as the reputation of your institution.

Compliance
Educational institutions are bound by laws and regulations like the Family Educational Rights
and Privacy Act (FERPA) in the U.S., the General Data Protection Regulation (GDPR) in the EU, and other
regional data protection laws. Adhering to these regulations while implementing AI systems is a complex
but necessary task.

Implement
Practical tips
• Ensure the data you are collecting and using is the minimum needed for the task. The less
data you collect, the less risk there is of a harmful data breach.

• Where possible, anonymize student data to protect student identities. This could involve

AI Navigators
removing personally identifiable information (PII) or replacing it with pseudonyms.

• Have a plan in place for responding to data breaches. This should include steps or identifying and
containing the breach. Research

Navigate back to
41 Microsoft Education AI Toolkit
table of contents
Engaging the community

Overview
AI technologies are revolutionizing education, presenting exciting opportunities for teaching and
learning. However, implementing AI-driven tools requires thoughtful planning, clear communication,
and collaboration with supportive communities. Stakeholders including administrators, educators,
students, parents, and community members have diverse responsibilities and seek understanding of
how AI tools will influence their daily experiences and future endeavors. This section addresses

significant challenges and opportunities in involving your community in the adoption of AI-powered
education tools. Additionally, it offers practical advice and strategies for:

• Building trust and support for AI-powered tools in education

Plan
• Understanding and addressing your community’s concerns

• Matching tools to your goals and needs

• Building a shared vision with your community

Building trust and support with stakeholders

Implement
Building trust with stakeholders across educational communities is a crucial aspect of any initiative.
There are many ways to build this trust:

• Communicate clearly and frequently about goals, expectations, and progress

• Seek feedback and input from the stakeholders

• Demonstrate competence and credibility by delivering high-quality results

A reliable method to establish trust involves aligning with shared objectives and values that prioritize

AI Navigators
student success by offering high-quality learning opportunities. Familiarizing yourself with the key
points below will empower you to engage in meaningful discussions with your community partners:

Key point

AI-powered tools help faculty and staff automate or streamline time-consuming tasks that interfere with
more crucial learning-related needs.

Email correspondence, exploring data trends, researching new instructional approaches, and drafting
detailed syllabi and course information take time away from connecting with and meeting diverse
students’ needs. Generative AI tools give educators time back so that they can refocus on what matters
Research

Navigate back to
42 Microsoft Education AI Toolkit
table of contents
most. Learn how educators in Wichita Public Schools used Copilot to become more efficient in the
customer stories in the AI Navigators tab.

Overview
Key point

AI-powered tools can also help address some of the critical challenges and opportunities
in education.

Accessibility is a key component of equitable schools. Generative AI tools can help educators create
high-interest text for emerging readers, develop multiple means of representation for content, and offer
new ways of demonstrating ideas for students. All of this adds up to a more engaging and equitable
experience for students. Read about how educators in Wichita Public Schools used Copilot to adapt
primary documents for their social studies classes in the customer stories in the AI Navigators tab.

Plan
Transformative workplace skills
Understanding how AI is already impacting the workplace, classrooms, and lecture halls, is critical to
preparing students and your community for adopting AI. As a fundamental component of the fourth
industrial revolution, AI, along with related fields such as machine learning and data analytics, is
transforming workplace skills and experiences.1 Medical research centers, various businesses, municipal
operations, and sustainable energy science are all driving rapid innovation. For instance, Walmart uses
AI technologies to streamline inventory management, ensuring the availability of the correct products

Implement
for their customers.2 Walmart also partnered with nearby Bentonville Schools to provide AI learning
experiences for local students.3

Your colleagues at schools and institutions have implemented a multi-tiered approach to address
evolving workplace needs. This includes the recent introduction of a K–12 vertical program that
integrates AI principles into every grade level and subject area.4 Across campuses nationwide, billions
of dollars have been allocated to develop programs, recruit faculty, and construct buildings to establish
new AI initiatives aimed at driving ongoing innovation.5

For instance, in early 2023, the University of Buffalo launched the National AI Institute for Exceptional

AI Navigators
Education.6 Their initial projects include the AI Screener, which identifies each student’s needs, and the AI
Orchestrator, which assists speech and language pathologists in creating personalized interventions.

Research

1
mckinsey.com/featured-insights/mckinsey-explainers/what-are-industry-4-0-the- 4
schools.gcpsk12.org/page/32147
fourth-industrial-revolution-and-4ir 5
insidehighered.com/news/tech-innovation/artificial-intelligence/2023/05/19/
2
tech.walmart.com/content/walmart-global-tech/en_us/news/articles/walmarts- colleges-race-hire-and-build-amid-ai-gold
ai-powered-inventory-system-brightens-the-holidays.html#:~:text=Our%20 6
buffalo.edu/ai4exceptionaled.html
AI%2Ddriven%20systems%20leverage,to%20differentiations%20by%20zip%20codes.
3
core-docs.s3.amazonaws.com/documents/asset/uploaded_file/3173/
Bentonville/2197839/News_Release_Ignite_Program_Expansion.pdf
Navigate back to
43 Microsoft Education AI Toolkit
table of contents
Understanding and addressing your community’s concerns

Overview
As you meet with different community members, you will encounter a multitude of concerns, interests,
and needs. Take advantage of this opportunity to cultivate empathy by actively listening to their
worries, demonstrating your AI knowledge and leadership expertise, and inspiring their support for your
AI initiative.

Leadership and administrators

Schools are increasingly the target of cyberattacks that target student data. Consequently, school
leaders are dedicating more thought, resources, and funding towards protecting user data.

Plan
“Student privacy is one of our biggest concerns especially when it comes
to AI tools. We are committed to vetting any tool before it’s introduced to
our classrooms to make sure that we know how our data is used and that,
ultimately, it’s safe and protected. We’re able to use Copilot for Security to
enhance and extend our ability to identify threats, automate our responses,
and remediate any issues.”

Implement
School leaders may have concerns about equity and accessibility when it comes to integrating AI.
Schools want to ensure that AI tools are accessible for all students, including those with disabilities,
and that the tools do not exacerbate existing inequalities.

“We will evaluate all AI tools to make sure they can be equitably accessed and

AI Navigators
used by students from various socio-economic backgrounds, different levels
of technology access and diverse learning needs. We aim to understand how
AIcan help us build a fairer educational landscape and remediate any issues, as
exemplified by institutions like the University of Texas.”
Research

Navigate back to
44 Microsoft Education AI Toolkit
table of contents
Educators and practitioners

Overview
Based on past experiences, educators may feel that new programs and initiatives are introduced,
supported for a short time, and then forgotten. Some teachers are hesitant to adopt technology
unless they are comfortable with their own skills and can support any questions or issues their
students may encounter.

“We are committed to making sure that you and your students know how
to use AI tools responsibly. Our plan includes age-appropriate instructional
materials, suggested conversation starters, guidance on modeling appropriate
use, and taking an iterative approach to adapting policies. You can also refer

Plan
to resources like Microsoft Learn’s Equip your students with AI and tech skills
for today—and tomorrow and Empower educators to explore the potential of
artificial intelligence courses for self-paced learning.”

Educators focus on the positive impact of instructional strategies and tools on learning. They are

Implement
generally more receptive to concepts or tools that are easy to adopt and have both immediate and
lasting effects.

“Early research indicates that students benefit from AI-generated explanations,

outperforming those who only receive correct answers.7 We also have Learning
Accelerators available that provide learners immediate, personalized coaching
to help develop foundational and workplace skills.”

AI Navigators
Research

7
papers.ssrn.com/sol3/papers.cfm?abstract_id=4641653

Navigate back to
45 Microsoft Education AI Toolkit
table of contents
Students and families

Overview
Families may have reservations about big corporations profiting from children's data and AI
inquiries. They may recall past instances of broken trust by companies and aim to shield their
children from similar experiences.

“We prioritize your student's privacy by thoroughly examining each company's

privacy policies. We ensure that their data practices align with our values,
prioritizing options that prioritize privacy and responsible data use. We believe
privacy is a fundamental right and only work with providers who share this belief.”

Plan
Families rely on schools to equip their children for their future aspirations and careers. They
anticipate that students will have access to the latest technologies and opportunities that pave
the way for a successful transition into adulthood.

Implement
“We've integrated AI features into the tools students use daily for learning,
creativity, and productivity. Additionally, we're exploring how other schools
have implemented AI guardrails. These guardrails help students access school-
specific chatbots designed to support their individual learning requirements.”

“Integrating AI tools into our instruction is part of our commitment to

preparing students for the future. Experts at the World Economic Forum
and McKinsey & Company have highlighted AI's significance in defining the
workplace.8,9 By incorporating AI into our teaching methods, we're ensuring that

AI Navigators
our students have the skills they need to thrive in this evolving landscape.”

Research

Navigate back to
46 Microsoft Education AI Toolkit
table of contents
The wider community

Overview
In addition to the previously mentioned concerns about having a high-performing school system
and ensuring student safety, the broader community expects their tax dollars to be used efficiently
and responsibly.

"We've been looking into how AI can enhance our data analysis efforts in
schools. What we've discovered is quite promising. With AI-powered tools, we
can analyze data in ways previously unimaginable. This means we can make
informed decisions on how to optimize our resources and staffing to best
support our students. Whether it's adjusting bus routes, optimizing utilities, or

Plan
refining staffing allocations, AI enables us to pinpoint areas for improvement
and redirect our focus and funds toward what truly matters: the learning
experience of our students."

AI tools have limitations and sometimes produce inaccurate information. We want our

Implement
students to graduate with solid knowledge and useful skills, so we should be careful about
how we use AI in our classrooms, if at all.

"We're excited to announce our plans to introduce age-appropriate, custom

chatbots for our students, inspired by successful initiatives like those in New
York City Public Schools. These chatbots will be tailored specifically for our
students, ensuring that the data used by the model comes from trusted
sources and that we have control over our users' data. This means our students
will have access to AI tools in a protected and safer environment compared
to public tools. We're committed to providing innovative yet secure learning

AI Navigators
experiences for our students."

Research

8
weforum.org/press/2020/01/the-reskilling-revolution-better-skills-better-jobs-better-education-for-a-billion-people-by-2030
9
mckinsey.com/featured-insights/mckinsey-explainers/what-are-industry-4-0-the-fourth-industrial-revolution-and-4ir

Navigate back to
47 Microsoft Education AI Toolkit
table of contents
 Continuing the conversation

Overview
No matter where you are in the process, you will continue to speak with a wide variety of
stakeholders—think educators, instructional leaders, school or college administrators, students,
families, and community members. Maintain a healthy relationship with these thought partners
with open dialogue.

How can I protect the privacy and security of students’ data when using AI-powered tools?
“We start by reviewing each AI tool’s terms of service and privacy policy to ensure that they
are committed to privacy and are aligned to our expectations. We know that Microsoft’s
generative AI solutions like Copilot, Copilot for Microsoft 365, and Azure AI Studio support

Plan
FERPA compliance and student data privacy protection. They use advanced encryption and
data handling policies to secure sensitive information. Microsoft’s AI solutions provide access
controls and transparency in data usage, undergoing regular compliance audits to maintain high
standards of privacy and security. We can customize privacy settings to align with our specific
compliance requirements and data governance policies.”

How do these AI solutions enhance student engagement and improve learning outcomes?
Are there studies or evidence demonstrating their effectiveness in educational settings?
“AI’s impact on instruction will keep evolving as practices, tools, and usage grow. Early findings
show that including generative AI explanations of concepts in instruction has a positive effect on

Implement
learning when compared with just giving corrective feedback.”10

How can I prevent academic dishonesty and plagiarism when using AI-powered tools?
“Protecting our school’s academic integrity begins with all users learning how to use AI
responsibly. We’re starting with professional development for educators, modeling responsible
use, and having open discussions. We’ve also paired our training with a clear policy

How can these AI solutions be customized to align with our district’s specific curriculum
standards, instructional goals, and educational objectives?
“We’re able to create customized AI-powered teaching assistants and configure data dashboards

AI Navigators
that help address our unique challenges. Even better, the data that we use in our schools is kept
private within the organization and is not used to train larger models.”

Research

10
papers.ssrn.com/sol3/papers.cfm?abstract_id=4641653

Navigate back to
48 Microsoft Education AI Toolkit
table of contents
 Overview
How can I support students with diverse learning needs and preferences when using AI-
powered tools?
“Our educators can use tools like Copilot and the Learning Accelerators to personalize
instructional content to meet individual student needs. With Copilot, educators can quickly
adapt content into different languages or reading levels. Furthermore, they can use prompts
to create custom explanations or analogies that build upon age-appropriate knowledge or a
student’s interests. Copilot supports multiple means of generating prompts including text- or
voice-based and Copilot for Microsoft 365 includes screen reading capabilities.”

Learn more at the Accessibility tools for Copilot for Microsoft 365 resource page.

Your vision, empathy, pedagogical prowess, and expert communication skills will help you

Plan
successfully implement AI-powered tools in your schools or institution. Continue to collaborate
with stakeholders and practice transparency as you address your community’s concerns, needs,
and goals.

Implement
AI Navigators
Research

Navigate back to
49 Microsoft Education AI Toolkit
table of contents
 Overview
Plan
Implement
Section 3
Implement

AI Navigators
Research

Navigate back to
table of contents
Getting started with Microsoft AI

Overview
Sitting down at a computer to learn how to use a generative AI tool is difficult when you are managing
a school district, running a technology department, or operating a university. There’s often too little
time and too many pressing issues for deep exploration and trial. This section provides instructions,
links, and additional resources to help you begin your AI journey.

Consider some of the responsibilities that leaders like yourself assume in your role and how generative
AI tools might help you save time or be more efficient.

Plan
Implement
Superintendent’s cabinet Provost’s office IT Department
Revise an acceptable use Analyze data and automatically Help educators troubleshoot
policy for secondary school identify insights in a reporting audio or video issues on
students to include emerging spreadsheet. their devices.
technologies like AI.

AI Navigators

Directors of Technology K-20 Ministries of Education

Gather IP addresses, ASNs, Understand key insights from
and known activities of malicious school reporting data.
websites that pose cybersecurity
risks for students and educators.
Research

Navigate back to
51 Microsoft Education AI Toolkit
table of contents
Microsoft Copilot tools and services

Overview
Microsoft Copilot helps you accomplish time-consuming tasks. For most of the tools, all that’s needed is
a basic understanding of how to get started. Write new content like policy frameworks, analyze reporting
data in spreadsheets, or conduct research on security threats.

This section provides instructions, links, and additional resources to help you begin your AI journey. Once
you know which Microsoft Copilot will help you with a task, just follow the steps and try it for yourself.

Plan
Microsoft Copilot Copilot for Microsoft 365
Find the right information, create Use content in Microsoft 365 applications
original content, and get things done like Word, PowerPoint, and Teams to
quickly without disclosing sensitive data. create new information, summarize,
and more.

Implement
Copilot in Windows Copilot for Security
Receive help, get answers, and jump Simplify incident response, threat
start projects within the Windows 11 hunting, intelligence gathering, and

AI Navigators
operating system. posture management.

Azure AI Studio
Build your own custom copilot using
GPT, DALL-E, and Whisper large
language models.
Research

Navigate back to
52 Microsoft Education AI Toolkit
table of contents
Microsoft Copilot

Overview
Microsoft Copilot is an AI assistant that helps leaders find
the right information, create original content, and complete
common tasks. It uses a chat interface and natural language
prompts to generate responses that summarize, analyze,
compare, and more. Copilot is available on the web and
mobile devices through Edge or other modern browsers.

When a user logs into Copilot using a free, school-issued Microsoft Copilot
Microsoft account, all prompts and responses are covered
by commercial data protection and Microsoft’s Customer
Copyright Commitment. These features help protect

Plan
sensitive data when included in chat prompts and increase Users
confidence in using responses from Copilot.
• Administrators
• Commercial data protection: Helps safeguard user and
organizational data because prompts and responses are • Staff
not saved when signed in with a school-issued
Microsoft account. Microsoft has no eyes-on access, and • Educators
chat data is not used to train the underlying model.
• Higher education students

Implement
• Customer Copyright Commitment Microsoft defends age 18 and older
customers from IP infringement claims arising from the
customer’s use and distribution of the output content
generated by Microsoft’s Copilot services or Azure
OpenAI Studio so long as the customer used the
guardrails and content filters we have built into our
products and implemented required mitigations.

AI Navigators
Research

Navigate back to
53 Microsoft Education AI Toolkit
table of contents
How to use Microsoft Copilot

Overview
1. Go to copilot.microsoft.com. Microsoft Copilot is also available as a shortcut in the Edge browser sidebar.

2. Sign in with a school-issued Microsoft account.

Plan
3.

Implement
Check for a green Protected badge in
Copilot. This badge indicates that commercial
data protection is enabled.

4. Select a conversational style. Precise is more

search-focused, Creative gives responses that
are more descriptive, and Balanced is

AI Navigators
somewhere in-between.

5. Create a prompt by following the Creating effective prompts section of this toolkit. Type the prompt
directly into the chat area, use your voice to speak your prompt, or add an image. Select the Send button.
Research

Navigate back to
54 Microsoft Education AI Toolkit
table of contents
6. Review Copilot’s response by checking the listed sources and reading for inaccuracies.
Copilot uses footnotes to cite sources within the response and includes links at the end

Overview
to check source material.

Plan
7. Copy or download Copilot’s response.

Implement
8. Select New Topic to clear Copilot and begin
a new prompt.

AI Navigators
Research

Navigate back to
55 Microsoft Education AI Toolkit
table of contents
Try Microsoft Copilot

Overview
Superintendent's cabinets or university administrators can quickly draft policy guidelines on generative
AI using Microsoft Copilot. Use the prompt below to see its capabilities..

Copilot prompt

Create a comprehensive policy on plagiarism and academic integrity specifically addressing the
use of artificial intelligence (AI) within our school district, which caters to a diverse audience of

Plan
20,000 students, their families, and educators. Write in a clear, direct tone that is accessible to a
general audience, including young students and individuals without a background in technology.
All sources of information and guidelines must be cited accurately and clearly within the document,
following APA citation style.

Implement
Copilot refinements Learn more
Add these refinements into the chat area to • Announcing Microsoft Copilot,
revise Microsoft Copilot’s response. your everyday AI companion

• Add a section about the Benefits • Expanding Microsoft Copilot access

of AI after the section on in education
Artificial Intelligence.

AI Navigators
• Adoption Kit
• Include information from
tech.ed.gov/ai/ in the • Documentation
Benefits of AI section. Research

Navigate back to
56 Microsoft Education AI Toolkit
table of contents
Copilot for Microsoft 365

Overview
Copilot for Microsoft 365 combines powerful large language
models to enhance productivity and amplify creativity with
Microsoft 365 apps. Like Microsoft Copilot with commercial
data protection, Copilot for Microsoft 365 uses natural
language prompts to easily complete tasks in popular
applications like Word, PowerPoint, Excel, Outlook, OneNote,
Microsoft Teams and more.
Copilot for Microsoft 365
Schools and institutions can purchase Copilot for Microsoft
365 as an add-on to their exisiting A3/A5 Microsoft 365

Plan
Education plan.
Users

• Administrators

• Staff

• Educators

• Students age 18 and older

Implement
AI Navigators
Research

Navigate back to
57 Microsoft Education AI Toolkit
table of contents
How to use Copilot for Microsoft 365

Overview
1. Go to office.com and sign in with a school-issued
Microsoft account. Copilot for Microsoft 365 is
only available after your organization has
purchased the product.

2. Open a Microsoft 365 app like Word, PowerPoint,

Excel, Outlook, or Teams. The steps in this guide
use PowerPoint to demonstrate Copilot for
Microsoft 365, so choose this app to follow along.

Plan
Implement
3. Select the Copilot button in the Home ribbon
of a new presentation.

4. Select Create a presentation from (file).

AI Navigators
Research

Navigate back to
58 Microsoft Education AI Toolkit
table of contents
5. Open a Word document that contains information
that you want to use in the presentation. Select

Overview
the Share button and choose the Copy
link option.

6. Return to PowerPoint and paste the document link

into Copilot. Select the Send button. Watch as
Copilot creates a presentation from the content

Plan
in your Word document!

Implement
AI Navigators
Research

Navigate back to
59 Microsoft Education AI Toolkit
table of contents
Try Copilot for Microsoft 365

Overview
A Provost’s office might want to create a PowerPoint presentation from board meeting notes to share
with deans and faculty members. By linking Copilot in PowerPoint to a Word file with meeting notes,
Copilot for Microsoft 365 generates a first draft that’s ready for editing and reviewing. If your school has
an add-on license, create a PowerPoint presentation from your own Word document or try a different
Microsoft 365 app with Copilot.

Copilot prompt

Plan
As the university's Provost, draft a concise email to the Technology Director expressing interest
in Microsoft 365 Copilot, emphasizing its collaborative and data analysis tools. Highlight the
potential for improved university workflows and productivity, especially in research and student
engagement. Propose a meeting with a Microsoft partner for tailored solutions, stressing the
urgency for the upcoming academic year.

Implement
Copilot refinements Learn more
Add these refinements into the chat area to
• How Microsoft Copilot for Microsoft 365 works
revise Copilot's response.
• Microsoft Copilot, a whole new way to work
• Include information from support
microsoft.com/copilot in the paragraph
• Copilot for work
about Copilot for Microsoft 365.

AI Navigators
• Adoption kit
• Add a postscript or PS at the end of the
email. Include a sentence about data
• Documentation
privacy and security for Copilot for
Microsoft 365 with citations.
Research

Navigate back to
60 Microsoft Education AI Toolkit
table of contents
Copilot in Windows

Overview
Copilot in Windows is an AI-powered assistant that provides
answers from across the web, supports creativity and
collaboration, and assists with tasks on Windows 11 devices. It
also offers a convenient way to change Windows settings and
launch troubleshooters when issues arise.

Copilot in Windows is built into updated versions of the

Windows 11 operating system and runs from the taskbar. Copilot in Windows
It is free to install on Windows devices and is protected
by commercial data protection with a school-issued
Microsoft account.

Plan
Users

• Administrators

• Staff

• Educators

Implement
AI Navigators
Research

Navigate back to
61 Microsoft Education AI Toolkit
table of contents
How to use Copilot in Windows

Overview
Sign into a Windows device using a school-issued Microsoft account. Copilot in Windows is not
available when signed into a device with a local account.

1. Select the Copilot button in the taskbar.

2. Select a conversational style. Precise is

more search-focused, Creative gives

Plan
responses that are more descriptive, and
Balanced is somewhere in-between.

3. Create a prompt by following the Creating

effective prompts section of this toolkit. Type
the prompt directly into the chat area, use your
voice to speak your prompt, or add an image.
Select the Send button.

Implement
4. If the prompt generates new content, check
for inaccuracies and review cited sources.
Otherwise, follow the instructions to perform
the requested task.

5. Select New Topic to clear Copilot and begin

a new prompt.

AI Navigators
Research

Navigate back to
62 Microsoft Education AI Toolkit
table of contents
Try Copilot in Windows

Overview
IT leaders or administrators often rely on multiple programs with many open applications when triaging
an incident. Copilot in Windows can help you organize open applications by typing the prompt, Create
snap layout.

Copilot prompt

As an IT administrator conducting a workshop for your colleagues at a large university, prepare a

Plan
guide titled “Top 5 Ways Copilot in Windows Boosts IT Staff Productivity”. For each point, provide
a brief explanation and an example of a Copilot in Windows command that can be used. The
command examples should be ready to use in Copilot in Windows.

Implement
Copilot refinements Learn more
Add these refinements into the chat area to
• Announcing Windows Copilot
revise Copilot in Windows response.
• Welcome to Copilot in Windows
• Include the same content but add a
section called FAQ which should
• Date and privacy
contain 5 questions and answers about

AI Navigators
Copilot in Windows.
• Documentation

• Include the same content but add an

additional question in the FAQ that
contains a question and answer about
data privacy and security with Copilot
in Windows.
Research

Navigate back to
63 Microsoft Education AI Toolkit
table of contents
Copilot for Security

Overview
Copilot for Security is an AI-powered security solution that
increases an IT administrator’s efficiency and capabilities. It
uses an assistive copilot experience and end-to-end scenarios
for incident response, threat hunting, intelligence gathering,
and posture management. It also responds to natural
language security questions and learns from user interactions
and enterprise preferences.

Schools can access Copilot for Security in two ways:

• Access Copilot for Security: Open Copilot for Security Copilot for Security

Plan
through a portal at securitycopilot.microsoft.com. This
portal connects licensed Microsoft products and
third-party services like Service Now so that Users
administrators can run queries that rely on security
signals from multiple products. • IT administrators

• Portal: Open Copilot for Security through a portal at

securitycopilot.microsoft.com. This portal connects
licensed Microsoft products and third-party services like

Implement
Service Now so that administrators can run queries that
rely on security signals from multiple products.

Copilot for Security is available for early access customers.

Schools must have a Microsoft Entra P1 or P2 license and
a Microsoft Defender for Endpoint P2 license. Note: The
steps in this section use the Copilot for Security portal to
demonstrate how to get started.

AI Navigators
Research

Navigate back to
64 Microsoft Education AI Toolkit
table of contents
How to use Copilot for Security

Overview
1. Navigate to securitycopilot.microsoft.com and log in with authorized credentials. you must be an early
access customer to log in.

2. Select the Manage plugins button in

the dashboard.

Plan
3. Toggle the switch for Defender
External Attack Surface Management,
Microsoft 365 Defender, and

Implement
Microsoft Defender Threat
Intelligence. The remaining steps in
this section use these plugins to
demonstrate Security Copilot
capabilities Licenses for each service
is required.

AI Navigators
Research

Navigate back to
65 Microsoft Education AI Toolkit
table of contents
4. Close the Manage plugins window.

Overview
5. Type a prompt into the chat area and press Send.

Plan
6. Security pilot provides relevant cited
information from its security-tuned LLM.

Implement
7. For additional ways to use Copilot for Security, explore Featured prompts for pre-created queries
that run common security tasks. Promptbooks run a series of prompts in sequence with each

AI Navigators
prompt building on the previous.

Research

Navigate back to
66 Microsoft Education AI Toolkit
table of contents
 Try Copilot for Security

Overview
Directors of Technology and IT administrators often investigate malicious websites that pose
cybersecurity risks. Copilot for Security provides critical information like IP addresses, ASNs, and what’s
known about websites using Microsoft’s global threat intelligence. Simply type, Tell me about the
INSERT_URL domain and Copilot for Security does the research for you.

Copilot prompt

Plan
As the Director of Technology managing a small school district’s IT staff, compile a desktop reference
guide featuring 10 Copilot for Security prompts. Each prompt should have a title, a brief explanation,
and a ready-to-use command example. Maintain a formal tone throughout the guide.

Implement
Copilot refinements Learn more
Add these refinements into the chat area to
• Microsoft Copilot for Security
revise Copilot for Security’s response.
• How Microsoft Copilot for Security works
• Include all the same information in the
desktop reference guide but also link
• Introducing Microsoft Copilot for
each prompt to the website where it

AI Navigators
Security for Empowering defenders at
came from.
the speed of AI

• Include all the same information in the

• Documentation
desktop reference guide but also
include 5 additional prompts focusing
• Onboarding
on security reporting at the end of the
guide. Link each prompt to its
source website.
Research

Navigate back to
67 Microsoft Education AI Toolkit
table of contents
Azure AI Studio

Overview
Azure AI Studio is a development environment that schools
and institutions can use to build custom Copilots from
popular LLM like GPT, DALL-E, and Whisper. With a custom
Copilot, schools can deploy an AI application that uses
internal data for a tuned chat experience.

Azure AI Studio is supported by Azure OpenAI Services so

that administrators can monitor undesirable inputs, outputs,
and misuse with their custom Copilots. Because Azure AI
Studio is a part of the Azure platform, deployments are Azure AI Studio
backed by Microsoft’s enterprise-grade security protection.

Plan
Customers with an Azure subscription can add Azure AI Users
Service rafter requesting access. Once activated, customers
can use the Azure AI Studio platform. Pricing follows a • IT administrators
pay-as-you-go (PAYG) structure and includes Provisioned
Throughout Units (PTUs) to ensure Copilots are ready for use.

Implement
AI Navigators
Research

Navigate back to
68 Microsoft Education AI Toolkit
table of contents
How to use Azure AI Studio

Overview
1. Apply for access to Azure AI Service and confirm that you have an active Azure subscription.

2. Open Azure AI Studio by going to oai.azure.com. Select the appropriate directory, Azure
subscription, and Azure AI resource after signing in with your access credentials.

3. Select Chat playground on the landing page.

The following steps cover how to build a custom
chat copilot.

Plan
4. In the Assistant setup panel, write a System
message or use a pre-built template that
users see when interacting with the copilot.
Select Add your data to include sources
that the copilot will reference when

Implement
generating responses.

AI Navigators
Research

Navigate back to
69 Microsoft Education AI Toolkit
table of contents
5. In the Configuration panel, use the Deployment
tab to choose the model for the copilot. The

Overview
Parameters tab includes settings for adjusting the
model's temperature, max length, top probabilities
and more.

Plan
6. Type test prompts in the User message field
to evaluate the copilot’s performance
before deployment.

Implement
7. Select the Deploy to button to create a new
web app or publishto an existing web app
when the copilot is ready for implementation.

8. Share the web address of the copilot with AI Navigators

people at your school or institution. Be sure
to review the Azure AI Studio documentation
for detailed instructions, troubleshooting
steps, and the FAQ about building a custom
copilot for your institution.
Research

Navigate back to
70 Microsoft Education AI Toolkit
table of contents
Try Azure AI Studio

Overview
Azure AI Studio simplify the process that a K-20 Ministry of Education would follow to build a custom
copilot for their schools. It enables IT administrators and developers to easily link school-specific data
sources with large language models for secure, tailored results.

Copilot prompt

As a K-20 education ministry leader supporting 50 schools, compose an email to school principals

Plan
about Azure AI Studio security and privacy. Explain its function, advantages over generic AI models,
and data privacy measures in everyday language with citations. Conclude by announcing upcoming AI
chat experiences.

Implement
Copilot refinements Learn more
Add these refinements into the chat area to • Fundamentals of Azure OpenAI Service
revise Microsoft Copilot’s response.
• Documentation for Azure OpenAI Service
• Include the same information in the
message but begin with a • Documentation for Azure AI Studio
customer story such as this one.

AI Navigators
• Data Privacy and Security
• Include all the same information in the
desktop reference guide but also
include 5 additional prompts focusing
on security reporting at the end of the
guide. Link each prompt to the website
where it came from.
Research

Navigate back to
71 Microsoft Education AI Toolkit
table of contents
Microsoft Copilot Capabilities

Overview
Different ways to interact with Copilot
Copilot is a generative AI tool capable of creating different types of content, such as text, images, music,
and code. It can process inputs in the form of text, images, or voice. You can get started by visiting
copilot.microsoft.com. It’s designed to work on various platforms, including Windows, mobile apps, and
the Edge browser, where it’s accessible via a sidebar. When used with Edge and Windows, Copilot has
additional enhanced features. This section introduces Copilot's features, access methods, and practical
applications to boost your creativity and streamline your workflows..

Plan
Voice interaction: In Windows, on the website, or in the mobile app, you can use your voice to
communicate with Copilot. Tap the microphone icon at the bottom of the screen, speak your query,
and submit it. This hands-free approach can be especially useful when you’re busy or multitasking.

Implement
Visual search: Upload photos, artwork, or other images, and Copilot will analyze them. This feature can
be handy for educators, designers, or any content creators in the educational sector.

Image generation: Leverage Copilot’s Designer image creator to generate logos, drawings, visual aids
or other images based on your text descriptions. Learn more by reviewing the image prompting

AI Navigators
101 guide.

Windows 11 integration: Access Copilot directly from your desktop with Windows 11 by clicking
the Copilot icon on the Taskbar. It opens in a sidebar on the right, allowing you to seamlessly interact
with Copilot, providing instant assistance without disrupting your tasks.
Research

Navigate back to
72 Microsoft Education AI Toolkit
table of contents
Mobile apps: You can stay connected with Copilot on-the-go. It is available as a mobile app for
both iOS/ iPadOS and Android. Download the app to your smartphone or tablet, and you’ll have

Overview
Copilot at your fingertips wherever you go

Optimizing Copilot in the Edge browser

Edge browser sidebar: If you’re using the Edge browser, there’s a unique integration that lets you
interact with Copilot through a sidebar. Simply launch Edge and click the Copilot icon in the upper
right corner. This integration facilitates real-time Copilot assistance while you navigate the web, perfect
for research and educational resource exploration.

Plan
Implement
AI Navigators
Research

Navigate back to
73 Microsoft Education AI Toolkit
table of contents
Use a screenshot: In the Edge browser sidebar you also have the option to take a screenshot and

Overview
have Copilot analyze it for you, describe the image, or pull key takeaways out of the image.

Plan
Insightful browsing: When using Copilot through the Edge browser sidebar, it can provide insights

Implement
on your current web page. Whether you’re researching, analyzing, or reading articles, Copilot can
enhance your browsing experience.

Copilot instructions
Open the Edge browser and navigate to an article, PDF, document, or website.

AI Navigators
Click the Copilot icon in the upper right corner.

Ask Copilot to:

• Provide the 3 top takeaways of an article, or

• Revise the document looking for gaps in knowledge or bias in language, or

• Summarize this pdf into key points and takeaways, or

• Ask a follow-up question to the topic of the article.

Research

Navigate back to
74 Microsoft Education AI Toolkit
table of contents
Creating effective prompts with

Overview
Microsoft Copilot
Microsoft Copilot can assist in content creation After navigating to your preferred access, make
and delivery in education. It is designed to sure to sign in with your work account. You will
support IT teams, educational leaders, and see a green protected tag next to your
educators by helping to increase productivity and credentials. This helps ensure your data is
collaboration. Copilot contributes by providing protected while you are using Copilot.
relevant answers and offering tailored solutions
to benefit the entire educational community.
Built-in safety

Plan
To make the best use of Microsoft Copilot, you
Microsoft prioritizes the protection of educational
need to be skilled in creating effective prompts
environments using Copilot, with commercial
that will help it to generate results that will help
data protection. User and organizational
you. Prompts are the messages or requests you
data are safeguarded, and chat prompts and
make of Copilot using the chat interface. As AI
responses in Copilot are not saved or accessed
continues to advance, the techniques for effective
by Microsoft for training purposes. Our Customer
prompting will also evolve—it’s an ongoing
Copyright Commitment further assures education
learning process. This section of the AI Toolkit will
customers that they can use our services and
help you learn the basics of creating high-quality
generated content with confidence about
prompts to get the best results.

Implement
copyright claims.

Accessing Precision prompting

Microsoft Copilot Generative AI models operate much like students
following instructions for a task. When you ask
There are four ways to access Copilot:
students to write an academic paper, they might
explore assorted topics. However, if you instruct
• Copilot.microsoft.com
them to write specifically about the impact of
child labor laws during the Industrial Revolution,

AI Navigators
• Microsoft Edge sidebar
you’ll likely receive more focused papers.
• Windows 11 taskbar*
This analogy extends to instructing generative AI.
Through a well-crafted prompt containing specific
• Microsoft Edge app on mobile
instructions or questions, users can guide the AI’s
responses towards desired outcomes, ensuring
*Copilot in Windows 11 may not be available on your PC
relevance, coherence, and accuracy. This not
only enhances the AI’s interpretability and utility
but also promotes a collaborative and efficient
interaction between humans and machines.
It unlocks the full potential of generative AI
Research

applications across various domains.

Navigate back to
75 Microsoft Education AI Toolkit
table of contents
Try it

Overview
Examine these examples to discern the differences between a poorly crafted prompt compared to a
well-crafted one.

First let’s look at this example of a poorly crafted prompt. This prompt is too vague. It lacks context,
specific topics and learning objectives, and types of activities. It is overly general and lacks lesson plan
elements. It links out to some websites for further exploration.

Create a 9th grade lesson plan for science

Plan
Implement
AI Navigators

Response from a poorly crafted prompt

Research

Navigate back to
76 Microsoft Education AI Toolkit
table of contents
Now try this prompt

Overview
The prompt below provides specific instructions, standard alignments, a topic, and desired components.
Notice the detail is much more specific to the topic, it linked to NGSS standards, created many elements
of a good lesson plan, and offered options. This provides the educator with a much more detailed plan
that they can then customize around their specific needs.

Copilot prompt

Plan
Design a detailed lesson plan for a 9th-grade biology class focusing on cellular respiration, aligned
with the Next Generation Science Standards (NGSS). Please structure the lesson into distinct segments:
a 10-minute warm-up activity to introduce the topic, a 20-minute engaging lecture with interactive
elements, a 30-minute hands-on activity to reinforce learning, and a 10-minute formative assessment
to gauge student understanding. Include in your plan: specific learning objectives, materials needed
for each segment, strategies for student engagement, differentiated instruction methods for diverse
learning styles, and clear assessment criteria based on the learning objectives. Ensure the lesson is
designed to be completed within a standard class period

Implement
Now add an additional refinement to your results.

• Please provide 5 different analogies to help students remember the 3 three stages of cell respiration.

• How might I make the lecture more interactive? Provide 3-5 ideas for this lesson.

• What are some scaffolds I could use with students that might be struggling with this content?

AI Navigators
Research

Navigate back to
77 Microsoft Education AI Toolkit
table of contents
Crafting effective prompts

Overview
Crafting clear and specific prompts is critical for receiving accurate and effective responses. The more
specific, and clear the prompt is the better the results will be for the user. This practice, known as
prompt engineering, ensures better outcomes when using generative AI.

Ask the tool to take on a role.

• Act as a mentor who provides personalized feedback, guidance, and
encouragement to educators.
Persona • You are a technology integration specialist who helps educators integrate
technology tools in the classroom.

Plan
• You are the leader of a K–12 district that wants to improve the quality, equity, and
efficiency of education.

What do you want the AI to do?

• Review and offer suggestions on improving a policy.
Objective • Outline a budget for the next school year.
• Create an action plan based on the minutes of a board meeting.

Implement
Specify who it’s for.
• Parents and guardians of our high school students
Audience • Elementary educators that teach art and music
• The president of the university

What does the tool need to know?

AI Navigators
• Background information
Context • Type of output (table, image, email, etc)
• Specific details related to the task

Set directions or limitations.

• In less than 500 words
Boundaries
• A friendly and courteous tone
• Make columns for x, y, and z
Research

Navigate back to
78 Microsoft Education AI Toolkit
table of contents
 Overview
Persona

Example Copilot prompt

You are a technology committee chair

tasked with evaluating online apps for Audience
Objective
student use. To assist high school students

Plan
with learning pronunciation, create a table
of online apps that can be downloaded
for teaching world languages.
Context
Include a column for each of the
following: the name of the tool, a brief Boundaries
summary, cost, and user ratings.

Implement
Copilot prompt

AI Navigators
You are a technology committee chair tasked with evaluating online apps for student use. To
assist high school students with learning pronunciation, create a table of online apps that can be
downloaded for teaching world languages. Include a column for each of the following: the name of
the tool, a brief summary, cost, and user ratings.
Research

Navigate back to
79 Microsoft Education AI Toolkit
table of contents
 Refine your prompt

Overview
Experimenting with various instructions, techniques, or word choices can yield different responses. If
the results don’t match your expectations or lack specificity, consider adjusting your prompt for better
outcomes. The key to refining AI responses to meet your unique needs is to explore different strategies.
Don’t be afraid to try different approaches until you find what works best for you!

Tips for getting better results

If you are struggling to get the results you need, refine your prompt using these suggestions.

Plan
Be clear Provide specific instructions about the task to be performed, explain the data
and specific context, and output requirements. Leave as little to interpretation as possible.

Use high quality and diverse examples to guide the AI to generate more
Give examples
relevant and accurate responses.

Implement
Be descriptive Use analogies and provide details.

Don’t use slang,

This may cause the AI to give low quality, inappropriate,
jargon, or
or unprofessional responses.
informal language

Don’t assume Don't assume the AI has knowledge outside of the given prompt.

AI Navigators
prior knowledge Always provide context and set expectations.

If you create a prompt that works well for one task, try using it as a template
and adjust it for similar tasks.
Re-purpose a Example: Design a lesson plan for a [course and level] that aligns with
successful prompt [standards] and concentrates on the topic of [topic]. The lesson should include
[list of required parts]. It should be structured [requirements].
Research

Navigate back to
80 Microsoft Education AI Toolkit
table of contents
Conversation styles

Overview
In addition to creating a better, more specific prompt, selecting the style in Copilot is important too.
There are 3 styles available in Copilot: Creative, Balanced, and Precise. They will yield different results
so you can try comparing results too!

Understanding Conversation Style in Copilot

Plan
Creative Balanced Precise

Original and Information and Concise and

imaginative chat friendly chat straightforward chat

Implement
Checking for accuracy
Remember AI is an assistant, not a replacement, for a human review and it can make mistakes, leading
to inaccurate or fabricated information. Always review your AI responses for accuracy, grammar,
and style. Additionally, verify the content created by AI is factual and check for any irrelevant or
inappropriate material.

AI Navigators
Research

Navigate back to
81 Microsoft Education AI Toolkit
table of contents
Interactive prompts

Overview
Try using the following prompts in Microsoft Copilot and then refine them to meet your needs. You can
also compare the results using different conversation styles.

Copilot Prompt

You are an academic administrator who works on revising policies. Develop a policy framework on
plagiarism and academic integrity when using AI for a university. This is an important topic because AI
tools can generate original and realistic content that can be used for academic purposes, but they can

Plan
also pose ethical and academic challenges, such as cheating, plagiarism, misrepresentation, falsification,
or fabrication of data. You need to consider how to prevent, detect, and address these issues and
promote ethical and responsible use of AI tools in academic contexts. Your policy framework should
include the following elements:

a. A definition of plagiarism and academic integrity with AI, and why they are
important for learning and education.

b. A list of the main challenges and opportunities of using AI tools for academic purposes, and how
they affect plagiarism and academic integrity.

Implement
c. A set of principles and guidelines for using AI tools ethically and responsibly in academic
contexts, and how to cite and credit AI-generated content.

d. A summary of the benefits and limitations of your policy framework, and how it can be
implemented and evaluated.

Your policy framework should be written in a formal and academic style, and it should be no longer than
1000 words. You should use reliable and credible sources to support your arguments and claims, and
you should cite them using the APA style.

AI Navigators
Copilot prompt

As a seasoned school administrator, create a feedback rubric for evaluating teachers’ instructional
methods This rubric, for use by fellow administrators, should feature four performance tiers,
encompass ten formative assessment aspects, align with the school’s ethos and standards, and
employ encouraging, growth-oriented language. It should exclude non-instructional elements and not
supersede ongoing professional development efforts.
Research

Navigate back to
82 Microsoft Education AI Toolkit
table of contents
 Overview
Copilot prompt Copilot prompt
As a cybersecurity expert, develop a
As a friendly and helpful instructional designer,
straightforward tutorial for school staff
your task is to assist teachers in explaining
on spotting phishing emails and social
the concepts of logos, pathos, and ethos to
engineering attacks. The aim is to improve
10th and 11th graders in an AP Language class.
staff’s ability to recognize phishing
The students have no prior knowledge of
attempts, minimize the risk of falling for
these concepts. Develop clear and simple yet
such attacks, and safeguard the school’s
detailed explanations, analogies, and examples

Plan
data. The tutorial should be applicable
for each concept. Remember, your goal is to
across various email platforms, focus on
make these concepts easily understandable
practical, non-technical measures, and
for the students.
include relevant examples.

Build your own prompt

Now that you are familiar with the elements used in building an effective prompt try building your own:

Implement
Persona Who would be considered the expert in this area?

Objective What is the task you want the AI to complete?

AI Navigators
Audience Who is this for?

Context What does the tool need to know to give the best results?

Boundaries What specific directions or limitations do you want to set on the tool?
Research

Navigate back to
83 Microsoft Education AI Toolkit
table of contents
Professional learning

Overview
Generative AI brings new technology and new learnings. A well-developed professional learning
plan for your community will facilitate informed adoption of generative AI, and promote responsible
practices, while ensuring your institution stays relevant and responsive to technological advancements.
For any tool adoption, the best place to start is low-stakes experimentation. We recommend you
practice with the Copilot prompts within all of the sections of this toolkit.

Conduct a needs assessment to identify the gaps and opportunities for

1 the adoption of generative AI in your institution. Consider the different

Plan
roles and levels of your learners and how they can benefit from AI skills.

Define clear and measurable objectives that align with your AI priorities
2 and vision. Think about what you want your learners to know and do after
completing the AI learning plan.

Implement
Select relevant and engaging content that covers the topics and skills
3 your learners need. You can use existing resources or create your own,
depending on your context and goals.

Choose the most appropriate delivery methods for your content, based
on your audience and resources. You can opt for in-person or online,
4 synchronous or asynchronous, or a combination of modes to suit your
learners’ preferences and needs. Note: If you are looking for support with
organizational professional development explore the Global Training Partners.

AI Navigators
Collect feedback and evaluate the effectiveness of your AI learning
plan. Use data and evidence to monitor the progress and outcomes of
5 your learners, and adjust your plan as needed to improve the quality and
impact of your AI education.

Foster a community of practice among your learners and educators,

where they can share their experiences, challenges, and best practices
6 with AI. Encourage them to continue learning and exploring new AI
technologies as they emerge.
Research

Navigate back to
84 Microsoft Education AI Toolkit
table of contents
Microsoft Learn

Overview
Microsoft offers a variety of free resources that can support you in designing and implementing your
AI learning experiences. Microsoft Learn provides technical documentation and self-paced professional
learning experiences for different roles and levels.

Microsoft Learn documentation and resources

Learning for K–12 educators For educators looking for

can be found at AI for resources to teach AI and
Track the training progress
Education where they can tech skills and provide
and certifications of your

Plan
discover the AI for education students with opportunity
learners within your tenant
learning path as well as to earn verified credentials
by using the Microsoft
the  Classroom toolkit where from Microsoft, we
Learn Organizational
they can unlock generative recommend the Microsoft
Reporting Overview.
AI safely and responsibly Learn for Educators Program
with students ages 13–15.    - Applied Skills.

Microsoft Learn self-paced professional learning experiences

Implement
MS-012 Prepare your Preparing for AI: The AI Preparing to use AI: How
organization for Copilot for learning journey for technical business leaders can build a
Microsoft 365 leaders foundation for AI success

This 3-module learning Gain essential knowledge to As you prepare to use AI,
experience examines the set up, deploy, and use AI discover the five pillars of AI
Microsoft Copilot for Microsoft solutions. Learn which features success. Check out this guide
365 design, as well as its security you need to enable to use or for business leaders, and help
and compliance features, build internal AI solutions.

AI Navigators
your organization thrive on its
and provides instruction on path to AI transformation. We
how to implement Copilot for Format: recommend this training for
Microsoft 365. Microsoft Learn collection business leaders and all learners.
Audience:
Format: Higher education IT leadership and Format:
Microsoft Learn learning path IT department; K-12 IT leadership
Microsoft Learn collection
Audience: and IT department
Higher education IT leadership and Audience:
IT department: K-12 IT leadership Higher education IT leadership and
and IT department IT department; K-12 IT leadership
and IT department
Research

Navigate back to
85 Microsoft Education AI Toolkit
table of contents
GitHub Education

Overview
GitHub is another location to explore professional learning experiences and learning paths. It offers free
developer tools, training, and support for students, teachers, and schools.

Artificial Intelligence for Beginners— Mastering GitHub Copilot for Paired Programming
A Curriculum
A 6 Lesson course teaching everything to know
A deep dive into Symbolic AI, Neural Networks, about harnessing GitHub Copilot and an AI
Computer Vision, Natural Language Processing, Paired Programing resource.
and more. Perfect for beginners, it covers

Plan
TensorFlow, PyTorch, and AI principles. Format: GitHub Education course
Audience: Higher education faculty, leadership,
Format: GitHub curriculum with hands-on IT leadership, and IT department
lessons, quizzes and labs
Audience: Higher education faculty and
students; Higher education IT department; K–12
IT department

Implement
Copilot prompt

As the IT Department Director, you're tasked with enhancing educational strategies through
technology. Design a detailed 1-hour professional development session for middle school
educators focused on integrating Microsoft Copilot or ChatGPT to improve student writing
across subjects. Specify:

AI Navigators
- Session Goal: Clarify the main objective.
- Learning Objectives: List specific skills or knowledge the educators will gain.
- Hands-On Activities: Detail interactive tasks involving ChatGPT, tailored to writing improvement.
- Materials Required: Itemize resources needed for the session.
- Ethical and Pedagogical Framework: Allocate time for discussing the responsible use of AI
in education.
- Evaluation Methods: Describe how educators' understanding and session effectiveness will
be assessed.
Ensure the plan is practical, directly applicable to classroom settings, and addresses educators'
current familiarity with AI tools.
Research

Navigate back to
86 Microsoft Education AI Toolkit
table of contents
 Overview
Plan
Section 4

Implement
AI Navigators

AI Navigators
Research

Navigate back to
table of contents
Education AI Navigators

Overview
Microsoft is excited to share the stories of institutions leading the way with research, experimentation,
testing, and deployment of generative AI solutions in education–the AI Navigators. These trailblazers
span a range of countries and educational organizations–from Ministries and State Departments of
Education to institutions of higher education as well as primary and secondary schools.

The AI Navigators are introduced in two ways: customer stories and technical implementation guides.
The customer stories describe how these early adopters applied Microsoft technologies to meet complex
challenges unique to the students and faculty they serve. They highlight common themes, including:

Plan
Data
AI-powered data insights uncover valuable trends and patterns that can enable
proactive decision-making and foster continuous improvement.

Equity and accessibility

AI-driven teaching solutions can prioritize equity for all learners by addressing barriers
to access.

Implement
Hyper-efficiency
AI-driven efficiency can streamline both teaching and administrative tasks to reduce
workload and enhance outcomes.

Technical implementation guides explain how these navigators implemented AI solutions from
Microsoft. These peer-to-peer stories reflect some of the very challenges you face today, and, hopefully,

AI Navigators
will inspire you to follow the trails they’ve blazed.

Both the customer stories and technical implementation guides provide suggestions, tips, and tricks.
Use the chart on the next two pages to learn more, read their customer stories, review the technical
implementation guides, or explore the AI solutions they are using.
Research

Navigate back to
88 Microsoft Education AI Toolkit
table of contents
 Microsoft Azure Other Microsoft Customer Implementation
Copilot AI Studio products story guide

Overview
K–12 ministries and state departments of education

Department for Education,

South Australia
• Learn more Get started

•
Washington State Office
Learn more
of the Superintendent of
Public Instruction

Plan
Sikshana Foundation
• Learn more

K–12 primary and secondary schools

Implement
Learn more Get started
New York City Public Schools

Wichita Public Schools

• Learn more Get started

Higher education institutions

• AI Navigators
California State University, Learn more Get started
San Marcos

University of South
Florida
• Learn more

Oregon State University

• Learn more
Research

Navigate back to
89 Microsoft Education AI Toolkit
table of contents
Customer stories

Overview
Microsoft’s Education AI Navigators are leading
the way in establishing best practices that
use generative AI to unlock the full potential
of educational data, foster inclusivity, and
streamline workflows.

Get to know some of these leading institutions

and how they are using Microsoft AI tools
to improve their organizations and make a
difference.

Plan
Implement
AI Navigators
Research

Navigate back to
90 Microsoft Education AI Toolkit
table of contents
Wichita Public Schools

Overview
With nearly 50,000 students and over 100 different languages spoken, the amount of time and energy
required of Wichita educators to individualize their lessons was becoming unsustainable. They needed a
solution that could bring diverse, tailored learning experiences into the classroom—swiftly and efficiently.

As existing Microsoft 365 A5 account users with Surface devices and Entra ID, the Wichita IT team
seamlessly led an early adoption program of Microsoft Copilot. Educators used generative AI capabilities
to increase their efficiency, quickly creating instructional materials that were accessible at different reading
levels and in different languages. They also found that they could generate authentic, project-based
learning experiences at different levels and streamline individualized student feedback on assignments.

Plan
How do your current needs align to the rationale behind Wichita’s story?
Is this implementation model a good fit for you?

What are the advantages of introducing Microsoft Copilot to faculty and staff?

What AI usage guidelines (privacy, data protection) must be in place before taking the
technical steps toward implementation?

Implement
Wichita educators use Copilot
to make learning more
accessible and bring a greater

AI Navigators
diversity of tailored learning
experiences to the classroom.

Equity and accessibility Hyper-efficiency Microsoft Copilot Implementation guide

Research

Navigate back to
91 Microsoft Education AI Toolkit
table of contents
New York Public Schools

Overview
As the largest public school system in the world, with more than 1 million students and 1,700 schools,
many NYC educators and district staff reported feeling overworked and overwhelmed. The district
needed a solution that could help reduce the workload while meeting the individual needs of students
and families.

District IT leaders partnered with Microsoft to create a data hub of close to 2 billion records, forming
the foundation for a custom-built AI teaching assistant and family communication tool with Azure AI
Studio. Educators used the AI assistant to scaffold feedback and help students discover answers on their
own, multiplying their ability to be several places at once. And when the custom chatbot was introduced
to the district website, call volume decreased by 30%—leading to reduced overhead costs and fewer
frustrated families.

Plan
How do your current needs align to the driving forces behind NYC’s story? Is this implementation
model a good fit?

What are the advantages of building your own custom AI application?

What district-level data management solutions must be in place before taking the first steps
toward building an AI chatbot?

Implement
NYC Schools develop a
custom AI-powered teaching
assistant to multiply teacher
effectiveness while also

AI Navigators
reducing parent and family
wait times at the district office.

Hyper-efficiency Data Azure AI Studio Implementation guide

Research

Navigate back to
92 Microsoft Education AI Toolkit
table of contents
Department for Education, South Australia

Overview
The Department for Education, South Australia is driven by a mission to equip their students for a future
where AI is everywhere. Leaders wanted to instill AI literacy and bring generative AI into classrooms, but
one question loomed large—how to do it responsibly?

IT leaders relied on Microsoft’s Azure AI Content Safety, an AI-powered platform that blocks
inappropriate input queries and filters any harmful responses. This allowed them to responsibly deploy
EdChat, a custom student-facing chatbot built with Azure AI Studio that is empowering students
with the skills they need to thrive in the era of AI. EdChat helps students find quick answers before
discussing more complex and nuanced questions with their teachers. Students are also learning how to
use AI prompts for feedback on their schoolwork, stimulating their creativity and critical thinking.

Plan
How do your current needs align to the driving forces behind South Australia’s AI story? Is this
implementation model a good fit?

What are the advantages of building your own custom AI application?

Does this model effectively address your stakeholders’ biggest concerns when it comes to
deploying AI safely and responsibly?

Implement
South Australian students are
supercharging their creativity
and critical thinking with AI in

AI Navigators
the classroom.

Equity and accessibility Hyper-efficiency Azure AI Studio Implementation guide

Research

Navigate back to
93 Microsoft Education AI Toolkit
table of contents
California State University, San Marcos

Overview
As a university with many first-generation students, CSUSM wanted to increase graduation rates and
empower social mobility for its diverse population. To do this, they knew they had to find a way to
connect with each student, personalize their college experience, and meet their individual needs.

CSUSM used Dynamics 365 Customer Insights “journeys” to tailor the faculty’s communications for
each student—both digitally and in person—while responding to students’ unique interactions and
preferences. Dynamics also transformed the school’s systems, which were fragmented and siloed, and
consolidated their data. University leaders used AI-powered insights to individualize communications
and points of interest for every student, resulting in greater attendance and engagement at school-
sponsored events and support that continued beyond graduation.

Plan
How do your current needs align to the driving forces behind CSUSM’s story? Is this
implementation model a good fit?`

What are the advantages of seeking insights into your students’ communication preferences?

Would this model effectively streamline your current data management systems?

Implement
University leaders use
Dynamics 365 and the
power of AI to establish a

AI Navigators
personalized connection with
every student.

Equity and accessibility Data Dynamics 365 Implementation guide

Research

Navigate back to
94 Microsoft Education AI Toolkit
table of contents
Washington State Office of the Superintendent

Overview
of Public Instruction
Education leaders in Washington state, led by Superintendent Chris Reykdal, are taking proactive steps
when it comes to AI use in schools. Washington is among the first states in the U.S. to publish official state-
level guidance on AI use in schools, including an implementation roadmap and guidelines for appropriate
AI usage for both staff and students.

Driving Washington’s AI roadmap is a central human-to-AI-to-human approach: “Start with human inquiry,
see what AI produces, and always close with human reflection, human edits, and human understanding of
what was produced.” This approach is also helping to drive the development of new teaching and learning
standards in ELA, Science, and Math that include AI as an embedded component of the curriculum, rather

Plan
than being siloed into a separate supplemental area. School leaders are confident that the new standards
will provide an opportunity for all students to develop the skills they’ll need to be ready for the world of
work with AI.

Implement
Washington leaders take
proactive steps toward AI
implementation with statewide
guidance and integrated

AI Navigators
AI teaching and learning
standards.

i
Equity and accessibility Human-centered AI guidance1
Research

1
ospi.k12.wa.us/sites/default/files/2024-01/human-centered-ai-guidance-k-12-public-schools.pdf

Navigate back to
95 Microsoft Education AI Toolkit
table of contents
University of South Florida

Overview
The University of South Florida, a pioneer in research and innovation, has taken a significant step
toward ensuring equity. They have become one of the first universities to provide all students, staff, and
faculty with equal access to technology by fully implementing the Microsoft 365 Copilot program. This
initiative underscores their commitment to creating an inclusive environment where everyone in the
organization has the same opportunities to thrive.

USF adopted a “platform approach” to AI that includes integrations with all areas of campus life, from
student and faculty research to administrative efficiencies. University leaders like CIO Sidney Fernandes
have found that “generative AI and the copilot we have started to use have shown us that a single
person can do much more work.” With this increased efficiency, faculty and students are finding
more time to work on meaningful research, while the IT service desk is focusing more time on solving

Plan
problems rather than performing triage.

Implement
University of South Florida
faculty and students adopt
Copilot for advanced research,

AI Navigators
data management, and
administrative efficiency.

Equity and accessibility Hyper-efficiency Microsoft Copilot

Research

Navigate back to
96 Microsoft Education AI Toolkit
table of contents
Oregon State University

Overview
Oregon State University is dedicated to conducting open and collaborative research while also
prioritizing the protection of sensitive data and upholding the institution's reputation. This delicate
balance requires a cybersecurity approach that is both robust and responsive.

Partnering with Microsoft, OSU was able to widely implement tools such as Copilot for Security,
Microsoft Sentinel, and Microsoft Defender quite rapidly. These tools helped the university to use
natural language to dialogue across security data to detect and respond to incidents rapidly, reducing
response times from weeks to mere minutes. It redefined their approach, shifting from a time-
consuming and reactive strategy to a more efficient and proactive one.

How do your current needs align to the driving forces behind OSU’s story?

Plan
What are the advantages of leveraging Copilot for Security to protect your students, staff, and their data?

Would this model effectively streamline your current cybersecurity and data management systems? .

Implement
“We once had the ability to
detect incidents in the timescale
of weeks. Now we detect things

AI Navigators
in matter of minutes.”
—David McMorries
Chief Information Security Officer at
Oregon State University

Hyper-efficiency Cybersecurity Copilot for Security Microsoft Sentinel Microsoft Defender Video
Research

Navigate back to
97 Microsoft Education AI Toolkit
table of contents
 Sikshana Foundation

Overview
India faces challenges such as larger class sizes (average teacher-student ratio of 1:33 versus 1:23 in other
countries) and educators managing multiple grades and subjects. The Sikshana Foundation aims to improve
education quality by focusing on the concept of "Shiksha," a Sanskrit term encompassing instruction, lessons,
learning, and the study of skills.

Understanding the time constraints faced by educators, Microsoft Research India has developed the Shiksha
copilot. This mobile-ready tool, powered by generative AI, assists educators in creating personalized learning
experiences, assignments, and activities.1 Importantly, it also lightens the workload for educators. The Shiksha
copilot, using the Azure OpenAI Service, seamlessly integrates educator insights with curriculum requirements
and learning objectives, thereby enhancing efficiency and effectiveness. It is designed to support multiple
languages and various input methods, making it accessible to a diverse range of users.�

Plan
How do your current needs align to the driving forces behind Shiksha Foundation’s story? Is this
implementation model a good fit?

What are the advantages of creating custom copilots to enhance personalization and alleviate workloads?

What AI usage guidelines (privacy, data protection) must be in place before taking the technical steps
toward implementation?

Implement
“Shiksha copilot is very easy to
use when compared to other
AI we have tried, because it is
mapped with our own syllabus

AI Navigators
and our own curriculum.”
—Gireesh K S, Teacher,
Government High School, Jalige

Hyper-efficiency Data Azure OpenAI Service Video

Research

1
timesofindia.indiatimes.com/gadgets-news/microsoft-develops-shiksha-copilot-to-help-indian-teachers-create-study-
material/articleshow/104922133.cms
� timesnownews.com/technology-science/how-microsofts-shiksha-copilot-is-helping-teachers-in-india-prepare-better-
lessons-article-107524567

Navigate back to
98 Microsoft Education AI Toolkit
table of contents
Technical

Overview
implementation
guides
Discover how Microsoft’s AI navigators are
pioneering AI implementation in education.

Follow in their footsteps to assess your

organization’s AI readiness, acquire the
necessary technology, and take the first steps
toward building your own AI capability while

Plan
using their implementation as your guide.

Implement
AI Navigators
Research

Navigate back to
99 Microsoft Education AI Toolkit
table of contents
 Technical implementation guide for

Overview
Wichita Public Schools

Preparing for success

Software and services
Ready to recreate Wichita Public Schools’ success? Consider the following
first steps:

Purchase the IT infrastructure, licenses, and software required to

introduce Copilot to your educators and staff
Entra ID

Plan
Find a cohort of school and instructional leaders who are enthusiastic
about developing both the AI program and opportunities for
professional learning

Communicate with stakeholders to fully understand their concerns

Microsoft Copilot when it comes to deploying AI safely and responsibly. Create a
plan that will address their concerns. Refer to the Engaging your
community section in this toolkit for additional information.

Familiarize yourself with a step-by-step technical overview of how

Implement
to recreate Wichita's success in the Implementation roadmap.
Edge browser

Read the full story

• Wichita Public Schools

personalized learning for
students using Microsoft Copilot “There is a highly documented anxiety “ping” that
affects teachers each Sunday evening. We wonder if we

AI Navigators
• Expanding Microsoft Copilot
access in education are ready for the coming week and if we have time to
get ready. When teachers embrace Microsoft Copilot
and begin to understand the time savings it represents,
I see the anxiety fade away, replaced by sighs of relief.”

—Dyane Smokorowski
Coordinator of Digital Literacy
Wichita Public Schools  
Research

Navigate back to
100 Microsoft Education AI Toolkit
table of contents
Implementation roadmap for Wichita Public Schools

Overview
Follow these steps to get started in building your own AI capability.

1. Verify or purchase a Microsoft 365 Education A1, A3, or A5 license.

2. Open Entra ID and select Single sign-on (SSO) from the sidebar. Follow the prompts to configure
the SAML protocol.

Plan
Implement
3. Open Powershell in the Azure portal and enable the Commercial data protection for Microsoft
Copilots service plan. The Copilot service plan is included with A1, A3, or A5 Microsoft 365
Education plans.

4. Optional: To require Copilot with commercial data protection for all users, update the DNS

AI Navigators
configuration by setting the DNS entry for www.bing.com to be a CNAME for nochat.bing.com.

5. Support educators with the Enhance teaching and learning with Microsoft Copilot course to help
write effective prompts.

Technical guides to follow

• Microsoft Copilot with commercial data protection • Microsoft Learn: Copilot commercial data protection

• Microsoft Copilot Adoption Kit • Microsoft Learn: Enhance teaching and learning
with Microsoft Copilot
Research

Navigate back to
101 Microsoft Education AI Toolkit
table of contents
 Technical implementation guide for

Overview
Department for Education South Australia

Preparing for success

Software and services
Ready to recreate South Australia’s success? Consider the following
first steps:

Purchase the IT infrastructure, licenses, and software required to

Azure AI Studio build your own custom AI application with Azure OpenAI Service.

Plan
Review the Overview of Responsible AI Practices and Azure OpenAI’s
Transparency Note for guidelines of responsible use of the service and
Read the full story
system limitations that may apply.

• Microsoft News: South Australian Find a cohort of school and instructional leaders eager to pilot an
students are supercharging their AI program.
creativity and critical thinking
with AI in the classroom Create a plan that fully addresses the concerns of all stakeholders,
including your community members. For support, refer to the

Implement
• Microsoft News: Learning in the Engaging your community section within the toolkit.
AI era; How South Australia’s
Department for Education is Familiarize yourself with a step-by-step technical overview of how to
empowering students and recreate South Australia's success in the Implementation roadmap.
teachers with AI in the classroom

• Microsoft News: How Azure AI

Content safety helps protect
users from the chatroom to the
I think that if we had buried our heads in the sand
classroom
and banned AI and chatbots in schools, students

AI Navigators
• EdChat—How schools are using would likely have continued using it at home to
the generative AI chatbot simply generate answers and churn out assignments.
By introducing it in schools as part of learning, we’re
ensuring that they really understand how it can
supercharge their thinking and creativity rather than
replace it.”
—Martin Westwell
Chief Executive of the SA
Department for Education
Research

Navigate back to
102 Microsoft Education AI Toolkit
table of contents
Implementation roadmap for Department for Education

Overview
South Australia
Follow these steps to get started in building your own AI capability.

1. Open Azure AI Studio. Explore the Playground features and use the assistant setup
tool to get started.

Plan
Implement
2. Build, test, and revise models in Azure AI Studio’s playgrounds. Follow the steps to
first deploy and test a chat model without your data. After the deployment is created,
select Open in playground.

AI Navigators
Research

Navigate back to
103 Microsoft Education AI Toolkit
table of contents
3. Select your deployed GPT chat model from 5. Once you’re satisfied with the experience
the Deployment drop-down menu and follow in Azure AI Studio, you can deploy your

Overview
the Assistant setup prompts to customize model as a standalone web application.
your chatbot. Follow the steps to deploy your web app,
using the assistant setup for guidance.

Plan
4. Follow the steps to add your own data to the
chatbot model. You can upload data files or 6. Configure your chatbot authentication and
select from data sets in Azure Blob Storage. privacy settings. Add an identity provider and
Note that Azure OpenAI needs both a storage create a new app registration.
resource and a search resource to access and
index your data. Your data is stored securely 7. Set up Azure AI Content Safety services and
within your Azure subscription. moderate the available features. Choose how
you want to moderate text content, moderate

Implement
image content, and monitor online activity
within your chatbot. The Quickstarts can guide
you through the technical details and generate
code for making service requests.

AI Navigators
Research

Navigate back to
104 Microsoft Education AI Toolkit
table of contents
8. Access the How-to guides for using the service in more specific or customized ways. Concepts will
provide in-depth explanations of Azure AI Content Safety service functionality and its features. You’ll

Overview
find useful tables that explain the different levels of severity, so you can make informed decisions about
which content you want to filter.

Plan
Implement
9. Finally, recruit a small but diverse cohort to pilot the chatbot program and provide feedback data.
Return to Azure AI Studio and Azure AI Content Safety to further refine the chatbot’s
features and functionality.

Technical guides to follow

AI Navigators
• Microsoft Learn: What is Azure • Microsoft Learn: Get started with
OpenAI Service? Azure Open AI Service

• Microsoft Learn: What is Azure AI • Microsoft Learn: Create a project in

Content Safety? Azure AI Studio

• Microsoft Learn: Fundamentals of

Azure OpenAI Service
Research

Navigate back to
105 Microsoft Education AI Toolkit
table of contents
 Technical implementation guide for

Overview
New York City Public Schools

Preparing for success

Software and service
Ready to recreate NYC Public Schools’ success? Consider the following
first steps:

Purchase the IT infrastructure, licenses, and software required to

introduce Azure AI Studio and to your school.
Azure AI Studio

Plan
Review the Overview of Responsible AI Practices and Azure OpenAI’s
Transparency Note for guidelines of responsible use of the service and
Read the full story system limitations that may apply.

Find a cohort of school and instructional leaders enthusiastic about

• Collaborating to bring AI
developing the AI program and professional learning opportunities.
innovation to education

Communicate with stakeholders to fully understand their concerns

• How NYC Public Schools
when it comes to deploying AI safely and responsibly. Create a plan
invited AI into its classrooms
that will address their concerns. Refer to the Engaging your community

Implement
section in the Plan tab for additional information.
• NYC Department of Education
Rises to the challenge with
Familiarize yourself with a step-by-step technical overview of how to
Microsoft Azure
recreate NYC Public Schools' success in the Implementation roadmap.

“Our mission is for students to graduate on a pathway

AI Navigators
to a rewarding career and long-term economic security,
equipped to be a positive force for change. If we are
not using AI in education, we’re putting our students at
risk of being behind.”
—Tara Carrozza
NYC Director of Digital
Learning Initiatives
Research

Navigate back to
106 Microsoft Education AI Toolkit
table of contents
Implementation roadmap for NYC Public Schools

Overview
Follow these steps to get started in building your own AI capability.

1. Explore the Playground.

Plan
Implement
2. Use the assistant setup.

AI Navigators
Research

Navigate back to
107 Microsoft Education AI Toolkit
table of contents
Build, test, and revise models in Azure AI Studio’s playgrounds

Overview
1. Build natural language prompts.

Plan
2. Apply prompt engineering.

Implement
AI Navigators
Research

Navigate back to
108 Microsoft Education AI Toolkit
table of contents
3. Generate code.

Overview
4. Generate images.

Plan
5. Use your own data.

Implement
6. Deploy your generative AI model.

7. Configure authentication and privacy settings.

Technical guides to follow

• Microsoft Learn: What is Azure AI Studio?

AI Navigators
• Microsoft Learn: Fundamentals of AzureOpenAI Service

• Microsoft Learn: Get started with Azure OpenAI Service

Research

Navigate back to
109 Microsoft Education AI Toolkit
table of contents
 Technical implementation guide for California

Overview
State University, San Marcos

Preparing for success

Software and services
Ready to recreate CSU San Marco’s success? Consider the following
first steps:

Purchase the IT infrastructure, licenses, and software required to

consolidate your current data management systems.
Dynamics 365 Customer

Plan
Insights with Copilot Work with stakeholders to come up with a comprehensive
communication plan that includes students, both prospective and
current, as well as faculty and staff.
Read the full story
Complete the Dynamics 365- Implementation Readiness Review to
proactively assess your institution’s readiness to start a Dynamics
• Microsoft Education Blog: 365
Streamline messaging with
Dynamics 365 Copilot Familiarize yourself with a step-by-step technical overview of

Implement
how to recreate CSU San Marco's success in the Implementation
• Microsoft Customer Stories: roadmap. implementation project.“U
CSUSM prioritizes the student
lifecycle journey with Dynamics
365 Customer Insights

Universities can be complicated for any student, but

AI Navigators
it can be especially challenging for first-generation
students. It’s important to know where each of our
students are in their lifecycle journeys.To do that, we
needed AI technologies that are flexible and can grow
with the university.”

—Tony Chung
Chief Information Officer
CSUSM
Research

Navigate back to
110 Microsoft Education AI Toolkit
table of contents
Implementation roadmap for California State University, San Marcos

Overview
Follow these steps to get started in building your own AI capability.

1. Address any categories from the Dynamics 365-Implementation Readiness Review

that influenced your results. The results of the assessment will provide actionable recommended
practices to prepare for a successful project start.

Plan
Implement
1. Log into the Dynamics 365 Implementation Portal to get started with building your project.

2. From the Projects tab, choose Create/Join Project. Follow the prompts for Project
Onboarding to Dynamics 365.

AI Navigators
Research

Navigate back to
111 Microsoft Education AI Toolkit
table of contents
3. Create a new project or join an existing one. Use the Implementation Wizard to capture
the scope, partner details, and go-live timeline.

Overview
4. Define your project characteristics and implementation details. From the Project Profile tab, manage
your project data and define your users.

Plan
Implement
5. Enhance your project with pre-built products and features for Dynamics 365. From the list of features,
choose Customer Journeys to unify your students’ data and gain AI-powered insights into their
communication preferences and online engagement.

AI Navigators
Research

Navigate back to
112 Microsoft Education AI Toolkit
table of contents
6. From the Implementation Guidance tab, filter and explore resources tailored to your project
needs. Read through the documentation and download relevant training materials.

Overview
Plan
7. Conduct project reviews with the Go-live Readiness Review tool. This will help identify any
potential risks prior to deployment. View the tool’s mitigation recommendations and best practices,
then make updates to the project as needed.

Implement
AI Navigators
8. Deploy the project. Use the Success by Design methodology for guidance in maintaining the long-
term health of your solution.

Technical guides to follow

• Get Started with Dynamics 365 Customer • Microsoft Learn: Work with Dynamics 365
Insights—Data Customer Insights Data

• Get Started with Customer Insights—Journeys • Microsoft Learn: Product overview for
Dynamics 365 Customer Insights
Research

Navigate back to
113 Microsoft Education AI Toolkit
table of contents
 Overview
Plan
Section 5

Implement
Research

AI Navigators
Research

Navigate back to
table of contents
Latest research and information

Overview
Research on the effective use and adoption of generative AI technologies has emerged as more and more
universities, schools, ministries and state departments of education have begun experimenting and adopting
them at various levels within their organization. The following collection includes links and short summaries.

Accelerate AI transformation with skill building: Microsoft report

March 2024 • Research

Plan
The report from Microsoft highlights a critical moment for businesses to invest in AI skill
building due to the rapid increase in artificial intelligence (AI) adoption. This report offers
statistics that point to critical shortage of skilled professionals, making talent scarcity the main barrier
to AI implementation at scale. The report recommends that companies develop a comprehensive AI
adoption strategy that includes a widespread skill-building initiative for all levels of employees. It offers
suggestions and resources for companies to undertake skill-building efforts.

Implement
AI in Education: Microsoft Special Report
March 2024 • Research

This report synthesizes the latest insights from Microsoft, partner organizations, and
academia on the opportunities and challenges of AI in education. Drawing on research
findings, the report focuses on four key areas: the need for clear communication and guidelines on AI
use, improving operational efficiency and productivity, potential benefits of AI for personalized learning,
and the skills students need to prepare for the future.

AI Navigators
Copilot Lab
2024 • Planning support

The Copilot Lab provides all the resources necessary to start using Copilot tools. It
includes introductory videos, overviews for getting the most out of Copilot in each
Microsoft app, examples of effective prompts, tips for better prompting, and information on how
Copilot protects your privacy. This page serves as a comprehensive guide to help you begin your
journey with Copilot tools.
Research

Navigate back to
115 Microsoft Education AI Toolkit
table of contents
 Microsoft New Future of Work Report

Overview
December 2023 • Research

In the past three years, there have been not one but two generational shifts in how
work gets done, both of which were only possible because of decades of research and
development. The first shift occurred when COVID made us realize how powerful remote and hybrid work
technologies had become, as well as how much science was available to guide us in how to (and how
not to) use these technologies. The second arrived this year, as it became clear that generative AI had
advanced to the point where it could be valuable to huge swaths of the work people do every day. This
report centers on the role of integrating generative AI large language models into the world of work.

Plan
AI and education: Guidance for policy-makers
2021 • Research

This report explores the significant potential of artificial Iitelligence (AI) in addressing key
challenges in education, innovating teaching methods, and accelerating progress toward
Sustainable Development Goal 4 (SDG 4). Despite the rapid technological advances, the report notes that

Implement
associated risks and challenges have outpaced policy debates and regulatory frameworks. It provides policy-
makers with guidance on leveraging AI opportunities while mitigating risks within the educational sector.

AI has the potential to disrupt education—but will we allow it to?

May 2023 • Thought leadership

Dr. Philippa Hardman is a scholar at the University of Cambridge and a thought leader

AI Navigators
in the world of education technology. In this Tedx Talk, she discusses the changes and
possibilities of AI in education and some of the resistance of education to be disrupted.

Research

Navigate back to
116 Microsoft Education AI Toolkit
table of contents
 All the unexpected ways ChatGPT is infiltrating students’ lives

Overview
June 2023 • Thought leadership

The article, by Susan Svrluga and Hannah Natanson, discusses the diverse ways students
are utilizing ChatGPT and other AI tools, extending beyond academic cheating to
include creative and practical applications like generating gift ideas, proofreading, and even providing
relationship advice. It highlights students’ experiences with AI in education, from aiding in homework
and writing to serving as a tool for mental health support. The piece also touches on the concerns
of AI’s impact on learning and future professions, with students and educators expressing both the
benefits and potential drawbacks of this technology. leveraging AI opportunities while mitigating risks
within the educational sector.

Plan
Artificial intelligence for human learning: A review of machine
learning techniques used in education research and a
suggestion of a learning design model
2024 • Research

Implement
This research paper explores AI and ML-based learning support systems and their impact on human
learning, focusing on previous techniques and proposing a framework for designing such systems. It
reviews existing research, discusses ML techniques, and suggests leveraging AI to enhance learning
support. The paper introduces a learning design model, Self-regulated Learning with AI Assistants
(SLAA), and categorizes AI and ML techniques into nine types, illustrating their potential applications.
It provides an overview of AI in education, addressing technological and pedagogical considerations,
challenges, and future research directions.

This paper highlights how AI (Artificial Intelligence) and ML (Machine Learning) can change and improve
educational methods. It gives a thorough explanation of different AI strategies and introduces the SLAA

AI Navigators
model. This serves as a guide for teachers and curriculum developers to use AI tools to make learning
more interactive, effective, and tailored to individual needs. this technology. leveraging AI opportunities
while mitigating risks within the educational sector.
Research

Navigate back to
117 Microsoft Education AI Toolkit
table of contents
 Australian Framework for Generative Artificial Intelligence
(AI) in Schools

Overview
November 2023 • Research

The Australian Framework for Generative AI in Schools seeks to guide the responsible and ethical use
of generative AI tools in ways that benefit students, schools, and society. The Framework supports
all people connected with school education including school leaders, teachers, support staff, service
providers, parents, guardians, students, and policy makers.

Plan
Building socially responsible conversational agents using big
data to support online learning: A case with Algebra Nation in
British Journal of Educational Technology
May 2022 • Research

This study explores the use of artificial intelligence (AI)-powered conversational agents (CAs) to increase
engagement in online student discussion forums, focusing on enhancing interaction while ensuring

Implement
discourse safety. While AI CAs can offer human-like, supportive responses, their potential for unsafe
content has been under-examined. The research introduces a safety-aware CA model specifically for
high school online algebra forums, comparing its performance with two leading models in terms of
safety and socio-emotional support.

For educational practitioners, the study suggests employing generation control, using open-sourced
models, and leveraging public APIs to monitor and enhance discourse safety. It also advises caution in
evaluating social support through automated means, as seemingly supportive texts may still be unsafe.
The results advocate for the implementation of safety-aware CAs to foster secure and supportive online
learning environments, education, the need for contextualized AI understanding, and how AI Readiness

AI Navigators
differs from general AI competency. An example of implementing AI Readiness at a higher education
institution is also provided.

Research

Navigate back to
118 Microsoft Education AI Toolkit
table of contents
 Designing Language Learning Experiences with Generative
AI Tools

Overview
(Chapter from book: AI in Language Teaching, Learning, and Assessment)

2024 • Planning support

The chapterm Designing Language Learning Experiences with Generative AI Tools, from the book AI
in Language Teaching, Learning and Assessment discusses the potential of artificial intelligence (AI),
particularly generative AI, in language education. It highlights how AI can help language learners
improve their skills, get immediate feedback, and customize their learning according to their levels
and needs. Not only students but also teachers can benefit from AI. Teachers can use AI to plan their
courses and lessons, create educational and evaluative materials, and meet the varying needs of their
students. However, the chapter points out a challenge: it's essential to design second language (L2)

Plan
learning environments well to fully benefit from AI while reducing possible negative effects. The chapter
promotes the idea that a well-thought-out approach, known as learning experience design (LXD), is
crucial for teachers incorporating AI into their instruction. It warns against random teaching strategies,
which can lead to inconsistent results, and advocates for the use of established LXD methods to
effectively utilize AI in language education.

Implement
Department of Education Office of Educational Technology:
Artificial Intelligence and the Future of Teaching and Learning
May 2023 • Planning support

The U.S. Department of Education Office of Educational Technology’s new policy report, Artificial
Intelligence and the Future of Teaching and Learning: Insights and Recommendations, addresses the
clear need for sharing knowledge, engaging educators, and refining technology plans and policies
for artificial intelligence (AI) use in education. The report describes AI as a rapidly-advancing set
of technologies for recognizing patterns in data and automating actions, and guides educators

AI Navigators
in understanding what these emerging technologies can do to advance educational goals—while
evaluating and limiting key risks.

Research

Navigate back to
119 Microsoft Education AI Toolkit
table of contents
 Do AI chatbots improve students learning outcomes?
Evidence from a meta-analysis

Overview
May 2023 • Research

In this 2023 British Journal of Educational Technology publication, researchers conducted a meta-
analysis of 24 randomized trials to assess the impact of AI chatbots on students’ learning outcomes. It
was found that AI chatbots significantly enhance learning outcomes, particularly in higher education
and during short-term interventions. This suggests that the novelty of AI chatbots can notably improve
learning in brief periods, though this effect diminishes over longer interventions, indicating the
importance of incorporating features like human-like avatars, gamification, and emotional intelligence
to sustain engagement and effectiveness.

Plan
.
Dr. Philippa Hardman
Thought leadership

Dr Phil's Newsletter, Powered by DOMS™️ AI Connecting the science of learning &

AI with the art of learning experience design. Read Dr. Phillippa's profile at www.linkedin.com/in/dr-

Implement
philippa-hardman-057851120/

EDUCAUSE QuickPoll Results: Adopting and Adapting to

Generative AI in Higher Ed Tech
April 2023 • Research

AI Navigators
The report provides initial insights into generative AI's emergence in higher education,
examining its increasing use and the potential long-term implications for staff and institutional
operations. The findings reveal that attitudes towards generative AI are becoming more positive, with a
growing integration into daily activities within educational institutions. Consequently, there is an urgent
need for these institutions to develop appropriate staffing and governance frameworks to manage the
adoption and application of generative AI technologies effectively.
Research

Navigate back to
120 Microsoft Education AI Toolkit
table of contents
 Embracing the future of Artificial Intelligence in the
classroom: the relevance of AI literacy, prompt engineering,

Overview
and critical thinking in modern education
February 2024 • Research

This discussion explores the profound influence of Artificial Intelligence (AI) in education, emphasizing
the importance of AI literacy, prompt engineering proficiency, and enhanced critical thinking abilities.
The introduction of AI in education represents a significant departure from traditional teaching
methods, offering tailored learning experiences and support for diverse educational needs, including
those of students with disabilities. However, this integration presents challenges, such as the necessity
for thorough educator training and curriculum adjustments to align with societal norms. AI literacy
is deemed essential, encompassing comprehension of AI technologies and their broader societal

Plan
implications. Prompt engineering is highlighted as a critical skill for eliciting specific responses from
AI systems, thereby enhancing educational experiences and fostering critical thinking. Strategies for
integrating these skills into educational curricula and pedagogical methods are analyzed, drawing
from a case study at a Swiss university and a review of relevant literature. Practical suggestions for
implementing AI in the classroom are provided.

Implement
Empowering educators to be AI-ready
June 2022 • Research

This paper introduces the concept of "AI Readiness" and a framework for AI Readiness
training. AI Readiness refers to helping people understand data-driven artificial
intelligence (AI) in a way that is tailored to their specific profession and workplace. Unlike general AI
education, AI Readiness training is contextual and participatory. It recognizes that different sectors
like law and education need AI knowledge contextualized to their unique needs, intricacies, and
sensitivities. The provided framework outlines 7 steps for achieving AI Readiness. This paper focuses on

AI Navigators
AI Readiness for the education and training sector. It discusses the current state of AI in education, the
need for contextualized AI understanding, and how AI Readiness differs from general AI competency.
An example of implementing AI Readiness at a higher education institution is also provided.
Research

Navigate back to
121 Microsoft Education AI Toolkit
table of contents
 Ethics of AI in Education: Towards a Community-Wide Framework

Overview
April 2021 • Research

While research in Artificial Intelligence in Education (AIED) aims to support student

learning, experience from other AI domains suggests good intentions alone are
insufficient. Explicitly considering issues like fairness, accountability, transparency, bias, autonomy,
agency, and inclusion is also needed. At a broader level, there is a need to differentiate between doing
ethical things and doing things ethically, understanding pedagogical choices ethically, and accounting
for unintended consequences. However, addressing these ethical questions is highly complex.

Plan
Examining Science Education in ChatGPT: An Exploratory Study
of Generative Artificial Intelligence
March 2023 • Research

This study explores three areas related to the generative AI ChatGPT in science education: (1) How
did ChatGPT answer questions related to science education? (2) What are some ways educators could
utilize ChatGPT in their science pedagogy? and (3) How has ChatGPT been utilized in this study, and

Implement
what are my reflections about its use as a research tool?

Educators must model responsible ChatGPT use, prioritize critical thinking, and set clear expectations.
ChatGPT could be a useful tool for designing science units, rubrics, and quizzes, but any AI-generated
resource requires critical evaluation and adaptation for specific teaching contexts. The study used
ChatGPT for editing assistance and narrative clarity experiments. The intention is to catalyze broader
conversation around using generative AI in science education.

Experimental Evidence on the Productivity Effects of

AI Navigators
Generative Artificial Intelligence
July 2023 • Research

This study explores the impact of ChatGPT on the productivity of mid-level writing tasks. College-
educated professionals were given incentivized writing tasks where half of them had access to ChatGPT.
For those with access to ChatGPT, productivity increased, taking users 40% less time to complete the task
and their quality increased 18%. Additionally, inequities decreased among the workers. The participants
exposed to ChatGPT during the study were significantly more likely to incorporate it into their work in the
weeks after the study.
Research

Navigate back to
122 Microsoft Education AI Toolkit
table of contents
 Finding High-Impact Opportunities for AI in Education

Overview
March 2024 • Research

International Data Corporation conducted a global study, sponsored by Microsoft, to

understand how K-12 and higher education institutions are approaching and benefiting
from AI implementation. Educational institutions are adopting AI in classrooms and administrative
settings to improve student satisfaction, enable faster innovation, and enhance faculty/staff productivity
and operational efficiency. This study includes data on the rise of AI in education, top use cases, the
need to advance AI strategies, challenges, and steps towards adoption.

Plan
Future-ready education: Empowering secondary school
students with digital skills
May 2023 • Research

The report from the Capgemini Research Institute emphasizes the urgent need for
enhanced digital literacy among students to navigate the complex, misinformation-filled online world
effectively. It identifies a significant gap in digital confidence, especially among secondary students in
rural areas, compared to their urban counterparts and the perceptions of their teachers and parents.

Implement
A key finding is the evolving role of artificial intelligence (AI) in education. While teachers recognize
the importance of integrating AI skills for future jobs, they are split on the use of AI tools like ChatGPT
in learning, weighing its educational benefits against potential risks. The report suggests that over half
of the teachers see the advantages of using ChatGPT outweighing the drawbacks and advocate for
updating curricula to include AI-generated content.

The report calls for comprehensive societal and corporate investment in digital education, highlighting the
necessity of making digital and media literacy a mandatory part of educational curricula and promoting a
collaborative approach among stakeholders to prepare students for the digital challenges of the workforce.

AI Navigators
Generative Artificial Intelligence for Education and Pedagogy
July 2023 • Research

From Cornell University: Educators are encouraged to integrate generative artificial

intelligence into their teaching strategies due to its prevalence in both current student
use and future workplaces, offering personalized learning and increased accessibility. However, AI also
poses challenges like potential cheating, biases, inaccuracies, and ethical issues. It proposes a flexible
framework considering academic integrity, accessibility, and privacy concerns.
Research

Navigate back to
123 Microsoft Education AI Toolkit
table of contents
 Guidance for generative AI in education and research

Overview
December 2023 • Planning support

UNESCO’s first global guidance on Generative AI in education to support countries to

implement immediate actions, plan long-term policies and develop human capacity to
ensure a human-centered vision of these new technologies.

How AI Revolutionizes Regional Language Education

March 2024 • Thought leadership

Plan
In "How AI Revolutionizes Regional Language Education," the authors delve into
the vital role of regional languages in fostering inclusivity and understanding within
educational settings. They explore how artificial intelligence (AI) can enhance language education,
offering practical strategies and real-world examples to demonstrate its transformative potential. From
AI-driven translation tools to personalized learning experiences, the book examines diverse ways in
which technology can create more accessible learning environments. However, it also acknowledges AI's
limitations, addressing ethical and practical considerations with sensitivity. Ultimately, the book aims
to inspire dialogue, collaboration, and innovation in regional language education, advocating for more

Implement
inclusive and equitable learning environments where linguistic diversity is embraced.

Impact of AI assistance on student agency

March 2024 • Research

AI-powered learning technologies are increasingly utilized for automating and

supporting learning activities, such as personalized reminders, real-time feedback, and

AI Navigators
study recommendations. While these technologies are generally perceived to benefit student learning,
their impact on student agency and self-regulation is not well understood. To address this gap, a
randomized controlled experiment was conducted with 1625 students across 10 courses, focusing on the
impact of AI assistance on student agency in peer feedback contexts. Results suggest that students tend
to rely on AI assistance rather than learning from it. Even when AI assistance is removed, self-regulated
strategies are less effective than AI. Hybrid approaches combining AI assistance with self-regulated
strategies did not show additional benefits. This study underscores the need to strike a balance between
AI assistance and fostering student agency to ensure active student participation and skill development.
As AI becomes more prevalent in education, exploring best practices and ethical considerations is vital
to maximize its potential while empowering learners.
Research

Navigate back to
124 Microsoft Education AI Toolkit
table of contents
 Learning Design to Support Student-AI Collaboration:
Perspectives of Leading Teachers for AI in Education

Overview
June 2022 • Research

This study explores teachers' perspectives on designing student-AI collaboration (SAC) in K-12
classrooms. Through interviews with 10 Korean leading teachers in AI in Education (AIED), the study
highlights the importance of capacity and subject-matter knowledge building as key learning goals
for SAC. Teachers emphasize interdisciplinary learning, authentic problem-solving, and creative tasks,
along with process-oriented and collaboration performance assessment, to facilitate SAC. They stress
the need for instruction on AI principles, data literacy, and AI ethics, while also advocating for AI to offer
instructional scaffolding and act as a learning mate to enhance student-AI interaction. Additionally,
teachers underscore the significance of systematic AIED policy, flexible school systems, and a culture

Plan
of collaborative learning. They anticipate students to progress through three stages of collaboration
with AI: learning about AI, learning from AI, and learning together. These insights have implications for
educational policies, AI design, and instructional design aimed at enhancing SAC in learning.

Math Education with Large Language Models: Perils or Promise?

December 2023 • Research

Implement
This study involved 1,200 participants to investigate how exposure to LLM-based
explanations affect learning. In the experiment’s learning phase, participants were given
practice problems with two differing key factors in a between-participants design. First, whether they
were required to attempt a problem before or after seeing the correct answer, and second, whether
participants were shown only the answer or were also exposed to an LLM-generated explanation of
the answer. Subsequently, all participants were tested on new test questions to assess how well they
had learned the underlying concepts. Overall, the study found that LLM-based explanations positively
impacted learning relative to seeing only correct answers.

AI Navigators
Research

Navigate back to
125 Microsoft Education AI Toolkit
table of contents
 MIT Technology Review: ChatGPT is going to change
education, not destroy it

Overview
April 6, 2023 • Planning support

This article discusses the evolving relationship between AI and education, initially
focusing on the educational community's swift response to ChatGPT, notably concerns regarding
academic integrity and the potential for student cheating. It progresses to reassess ChatGPT's utility
as an interactive teaching aid that could substantially improve learning experiences. Addressing
the challenges educators face in adopting such new technologies, the article acknowledges the
development of innovative strategies to navigate these difficulties. Concluding, it considers the long-
term effects of AI in education, emphasizing the need for careful integration and the promise of
personalized learning experiences facilitated by advancements like ChatGPT.

Plan
National 4-H Council Youth AI Survey
February 2024 • Research

This resource is a national survey of 1,510 young people ages 9-17 by Hart Research and
supported by Microsoft. The survey shows that while most young people (66%) express

Implement
at least some understanding of what generative AI is and how it can be used, many kids (72%) are also
seeking support from adults in learning how to use these tools correctly and with confidence. Findings
conclude kids as young as age 9 have a solid understanding of generative AI tools and concepts and the
adults involved in their learning to also understand and engage with GenAI tools.

One Useful Thing

2024 • Thought leadership

AI Navigators
Ethan Mollick is an Associate Professor of Management at the Wharton School of the
University of Pennsylvania who studies entrepreneurship, innovation, and AI. His work on
One Useful Thing explores how he and his students are using AI tools in the school of business and in
entrepreneurial opportunities. He has published numerous works on AI including Co-Intelligence: Living
and Working with AI.
Research

Navigate back to
126 Microsoft Education AI Toolkit
table of contents
 Oxford University Press: AI in education: where we are and
what happens next

Overview
October 2023 • Planning support

This report synthesizes insights from a global network of experts and research to provide a
comprehensive overview of the current state of AI in education, its impact on teachers and students,
and the widening digital divide. It examines AI's varying role in education worldwide, teacher
perceptions and readiness for AI integration, and the influence of AI on student learning, emphasizing
the need for appropriate skills and safeguards against misinformation.

Key issues include the potential of AI as a tool for equality or a factor in widening the digital
divide, given the disparities in internet and AI access. Finally, the report proposes five strategic

Plan
recommendations for school leaders, education bodies, and governments to ensure that educational
advancements drive technological adoption rather than being dictated by it, aiming to create a more
inclusive, informed, and prepared learning environment.

Revealing an AI Literacy Framework for Learners and Educators

February 2024 • Planning support

Implement
A framework developed by Digital Promise that emphasizes that understanding and
evaluating AI are critical to making informed decisions about if and how to use AI in
learning environments. Recently, the framework has been expanded to support learners, teachers,
education leaders, and caregivers with the knowledge and resources they need to understand, use, and
evaluate AI.

AI Navigators
Research

Navigate back to
127 Microsoft Education AI Toolkit
table of contents
 Revolutionizing education: Artificial intelligence empowered
learning in higher education

Overview
December 2023 • Research

This study examines how global businesses' embrace of Industry 4.0 and digital transformation parallels
the educational sector's use of digital tools for personalized learning and equality. It specifically explores
the impact of artificial intelligence (AI)-based tools in education, focusing on the awareness and adoption
of AI among faculty members in higher education. The research investigates how AI has enhanced
learning experiences and influenced faculty work engagement.

The results indicate that implementing AI in education leads to the development of more effective
evaluation and assessment methods, thereby increasing faculty engagement levels. Key factors

Plan
identified in influencing the adoption of AI and work engagement include perceived risks, performance
expectations, and the level of awareness about AI, with attitude and behavior serving as crucial
mediating variables. This study highlights the significant role of AI in shaping the future of higher
education and underscores the importance of faculty awareness and attitudes toward AI adoption.
generative AI in science education.

Student-ChatGPT Dialogue in EFL Writing

Implement
September 2023 • Research

This study is based on a semester-long longitudinal experiment involving 212 college

students enrolled in English as Foreign Language (EFL) writing courses using generative
AI to revise essays. It analyzes students’ usage patterns and perceptions regarding generative AI with
respect to their intent and satisfaction.

Supporting students’ self-regulated learning in online

learning using artificial intelligence applications

AI Navigators
June 2023 • Research

Self-regulated learning (SRL) plays a pivotal role in aiding students in achieving high academic performance
and meeting their learning objectives, particularly in online learning environments. However, learners
often encounter obstacles in effectively implementing SRL strategies online. While recent advancements
in artificial intelligence (AI) hold promise in supporting learners' self-regulation online, research in this area
remains nascent. This study aims to investigate students' perceptions regarding the use of AI applications
to bolster SRL and to pinpoint the pedagogical and psychological factors crucial for their effective
utilization. Results indicated that learners found AI applications valuable for supporting metacognitive,
cognitive, and behavioral regulation across various SRL domains but not for regulating motivation.
Research

Navigate back to
128 Microsoft Education AI Toolkit
table of contents
 Teaching C550 with AI: Leveraging Generative Artificial
Intelligence in Computer Science Education

Overview
February 2024 • Research

This paper outlines how AI tools have improved teaching and learning within Harvard’s CS50 by explaining
code snippets, improving code style, and responding accurately to curricular and administrative queries
on the course’s discussion forum. A suite set of AI-based tools was integrated into the CS50 course.
Initially piloted with 70 students during the summer, these tools expanded to thousands of online students
and several hundred on campus in fall 2023. The course policy promoted these tools over commercial
AI software, like ChatGPT or GitHub Copilot, fostering guidance toward solutions rather than direct
answers. Students responded positively, likening the experience to having a personal tutor. Their findings
demonstrate how integrating AI thoughtfully in education can enhance the learners’ experiences and

Plan
provide customized support while enabling educators to address more complex pedagogical issues.

The Cool Cat Teacher Blog

2024 • Thought leadership

The Cool Cat Teacher Blog by Vicki Davis provides how-to guides, insights, and practical

Implement
strategies to incorporate generative AI tools into K–12 education. Examples of this work
include using AI to create substitute lesson plans, building your own GPT’s, and ways to
develop automations using AI that save time and make teachers more efficient.

The effectiveness of artificial intelligence on learning

achievement and learning perception: A meta-analysis

December 2021 • Research

AI Navigators
research undertakes a comprehensive meta-analysis to evaluate the impact of artificial
intelligence (AI) technologies on learning outcomes, specifically focusing on learning achievement and
learning perception within the educational field, an area previously lacking in quantitative analysis. This
analysis shows that AI technologies notably improve learning achievement but have a lesser impact
on learning perception. The effectiveness of AI in education is significantly influenced by thirteen
moderating variables, including sample demographics, learning content nature, methodological
approaches, educational settings, intervention lengths, and AI applications and technologies. These
findings underscore the importance of customizing AI solutions in education to account for various
contextual and technical factors, offering essential insights for educators and practitioners to enhance
learning outcomes and perceptions.
Research

Navigate back to
129 Microsoft Education AI Toolkit
table of contents
 The era of generative AI: transforming academic libraries,
education, and research

Overview
Chapter from book: Empowering Minds : Collaborative Learning Platform for Teachers,
Librarians and Researchers

January 2024 • Research

In this excerpted chapter from "Empowering Minds: Collaborative Learning Platform for Teachers,
Librarians, and Researchers," the authors delve into the transformative impact of generative artificial
intelligence (AI) on libraries, education, and research. They highlight how AI's ability to create novel
content and innovative solutions revolutionizes information processing and dissemination across various
sectors. Generative AI transcends traditional roles, ushering in an era where human-AI collaboration
blurs the line between technological aid and human inventiveness. Libraries are reimagined as dynamic

Plan
hubs of knowledge, offering enriched user experiences and more efficient resource management,
while educational settings benefit from tailored learning experiences that adapt to individual needs.
In academic research, generative AI streamlines traditional methodologies, accelerating the research
process and fostering interdisciplinary collaboration. The chapter critically examines the advantages of
generative AI, such as enhanced information accessibility and personalized learning experiences, while
also addressing ethical considerations and limitations. It emphasizes the importance of navigating
challenges to fully leverage AI's potential in a rapidly evolving digital landscape.

Implement
To Ban or Embrace: Students’ Perceptions Towards Adopting
Advanced AI Chatbots in Schools
October 2023 • Research

The paper examines the academic community's varied reactions to the introduction of
ChatGPT, noting that while there have been many discussions, student perspectives are often missing.
The study focuses on student reactions to a New York Times article on ChatGPT, applying the Theory

AI Navigators
of Reasoned Action to understand their views. It looks at how students' beliefs, personal experiences,
and societal expectations shape their opinions on whether ChatGPT should be used or banned in U.S.
educational settings. The findings reveal that while students see ChatGPT as a beneficial educational
tool, they also express concerns about its potential to enable cheating, spread misinformation, and
create equity issues.
Research

Navigate back to
130 Microsoft Education AI Toolkit
table of contents
 Transmission Versus Truth, Imitation Versus Innovation:
What Children Can Do That Large Language and Language-

Overview
and-Vision Models Cannot (Yet)
October 2023 • Research

The research explores the role of advanced artificial intelligence (AI) models, such as large language
models, positioning them not merely as intelligent entities but as cultural tools that facilitate the spread
of knowledge. The study investigates the capacity of these AI systems to replicate human learning
processes and assesses their ability to generate new ideas or understand cause and effect, akin to the
abilities of human children. It initiates the process of identifying which skills or types of knowledge can
be acquired by AI through various learning methods and datasets. Specifically, it delves into the kinds
of learning abilities that can be enhanced through the statistical analysis of extensive linguistic data. The

Plan
findings indicate that these machines may require more than substantial language and visual data to
reach the levels of creativity and innovation demonstrated by young children.

Unlocking the Power of ChatGPT: A Framework

for Applying Generative AI in Education

Implement
April 2023 • Research

This study examines the impact of AI chatbots like ChatGPT on education, highlighting the potential
for personalized learning experiences and streamlined feedback processes. It also acknowledges
significant hurdles, including the unproven efficacy of these technologies, data quality issues, and ethical
considerations.

What Makes Problem-Solving Practice Effective? Comparing

Paper and AI Tutoring

AI Navigators
August 2023 • Research

his study compares learning outcomes between using an intelligent tutoring system
(ITS) and traditional paper-based problem-solving for middle school students practicing linear graphs.
It finds that students using the ITS had over twice the eventually correct practice opportunities. These
findings suggest that while tutoring allows students to navigate challenging steps with tutor assistance,
eventually correct opportunities may not be the sole driver of learning gains. This study offers
quantitative evidence that ITSs provide more learning opportunities than paper-and-pencil practice,
highlighting the importance of effective instruction in maximizing learning gains.
Research

Navigate back to
131 Microsoft Education AI Toolkit
table of contents
 What’s New in Copilot for Microsoft 365

Overview
February 2024 • Planning support

This monthly blog brings you the latest updates, features and more information to help
you get the most out of your Copilot Microsoft 365 experience.

Plan
Implement
AI Navigators
Research

Navigate back to
132 Microsoft Education AI Toolkit
table of contents
Frequently asked questions

Overview
What is generative AI?
Generative AI refers to artificial intelligence technologies that can create content, such as text,
images, or simulations, by learning from vast amounts of data.

Plan
Which Microsoft tools can be used by education institutions?
All Microsoft tools can be used by education institutions and some generative AI tools are
restricted by age. Depending on the age of the user, education institutions can leverage various
Microsoft tools, including Microsoft Copilot, Copilot for Microsoft 365, Copilot in Windows,
Microsoft Teams for Education, Azure AI for custom solutions, GitHub for Education, and Learning
Accelerators. Learn more about all of Microsoft’s AI solutions.

Are there studies or evidence demonstrating the effectiveness of AI tools in

Implement
educational settings?
Yes, there are studies indicating that AI tools can significantly benefit educational settings by
providing personalized learning experiences, automating administrative tasks, and facilitating data-
driven decision-making, thereby improving learning outcomes and operational efficiency. One
example published in December 2023, Math Education with Large Language Models1, found that
generative AI-based math instruction positively impacted learning.

How do Microsoft’s AI solutions support accessibility for all students, including those with
disabilities?
Microsoft incorporates features like speech-to-text, language translation, content readers, voice

AI Navigators
assistances, computer vision and personalized learning options, which can help to accommodate
diverse abilities and learning needs, including those of students with disabilities.

What mechanisms are in place to safeguard that the content generated by these AI solutions
is appropriate for school settings and safe for students?
Microsoft’s AI solutions are governed by guidelines to ensure age-appropriateness and help
safeguard a positive educational environment. Read more about Microsoft’s commitment to
Responsible AI.
Research

1
Math Education with Large Language Models

Navigate back to
133 Microsoft Education AI Toolkit
table of contents
 How can I evaluate student work that uses generative AI?

Overview
To evaluate student work involving generative AI, establish criteria that focus on critical
thinking, problem-solving, and creativity, while ensuring students understand and articulate
AI’s role in their process. For example, you can assess their ability to critically evaluate
the AI-generated content as compared to non-AI content as part of lit reviews and other
content analysis.

How can I protect the privacy and security of students’ data when using
AI-powered tools?
Protect student data privacy by using AI tools that comply with data protection laws,
implementing robust cybersecurity measures, and educating students on digital privacy.

Plan
How can I prevent academic dishonesty and plagiarism when using AI-powered tools?
Help prevent academic dishonesty by setting clear expectations and encouraging originality
and critical engagement with AI-generated content. Consider updating your course
syllabus to set clear expectations about AI use.

How can Microsoft AI solutions be customized to align with our institution or

department’s specific curriculum standards, instructional goals, and educational

Implement
objectives?
AI solutions can be tailored to district curricula and aligned with standards and goals,
through customizable content and adaptable learning modules. Review the education
prompts library on GitHub for examples of how AI can help you do this.

Do these AI solutions provide analytics and reporting features that can help educators
track student progress, identify learning gaps, and inform instructional decisions?
AI solutions often include analytics and reporting tools, enabling educators to monitor
student performance, pinpoint learning gaps, and guide instructional strategies.

AI Navigators
Customizations using Azure OpenAI Studio and Microsoft Fabric can help provide detailed
insights and analysis. See the AI Navigators to review examples.

What kind of training and professional development will be provided for educators
and staff to effectively use and integrate these AI solutions into their instructional
practices?
Training on the effective use of AI tools for faculty and staff should include integration
strategies, pedagogical approaches, and ongoing support to ensure successful adoption. See
the professional learning section for additional information.
Research

Navigate back to
134 Microsoft Education AI Toolkit
table of contents
 Overview
Plan
Implement
AI Navigators

Microsoft and the Microsoft logo are trademarks of Microsoft or its subsidiaries. *Other names and
Research

brands may be claimed as the property of others.

Copyright © 2024 Microsoft