UNIT 1: REVISITING AI PROJECT CYCLE AND ETHICAL FRAMEWORKS FOR AI

Subjective Type Questions

A. Short answer type questions.
1. Why is it important for the data used in AI projects to be authentic and reliable?

Authentic and reliable data is crucial because it forms the basis of AI projects, ensuring accurate
results and avoiding potential conflicts.

2. Why are ethical frameworks important in AI development?

Ethical frameworks are important in AI development to ensure that AI systems are designed and
used responsibly, preventing harm, promoting fairness, ensuring transparency, and respecting
privacy, which builds trust and ensures societal benefit.

3. What is the role of accountability in AI ethics?

Accountability ensures that humans are responsible for the actions and outcomes of AI systems, and
corrective actions are taken if harm is caused.

4. Why is bioethics important in healthcare?

Bioethics ensures that medical practices and technologies respect human dignity, protect patient
rights, promote fairness, and prevent harm. It guides healthcare professionals in making responsible
decisions that consider the well-being of patients and society.

5. How does natural language processing (NLP) work in AI?

NLP is a domain of AI that enables machines to understand, interpret, and generate human language.
It is used in applications like speech recognition, chatbots, and language translation.

6. What is the purpose of data exploration in the AI Project Cycle?

Data exploration helps in understanding and analyzing the collected data by identifying patterns,
trends, relationships, and anomalies. It ensures that the data is meaningful, complete, and useful
for building an accurate AI model.

7. What role does Natural Language Processing (NLP) play in AI?

NLP allows machines to understand, interpret, and respond to human language. It is used in virtual
assistants (like Siri), chatbots, language translation, and voice commands, making interaction with
machines more natural.

8. What is the importance of fairness in AI ethics?

Fairness in AI ensures that AI systems do not show bias or discrimination towards any individual or
group. It promotes equal and just treatment, avoids stereotyping, and builds trust in AI systems.
 9. How does transparency influence AI system development?

Transparency in AI means that the working of AI models is understandable and explainable. It builds
user trust, ensures accountability, and helps in detecting errors or biases in decision-making
processes.

10. What are the types of data used in AI projects?

1. Structured data: It is organised in a table within rows and columns, making it easy to analyse and
process data. It includes information such as name, address, quantity, stock price, etc. Data stored
in databases and spreadsheets comes under the examples of structured data.

2. Unstructured data: It does not have a specific format to store data, making it more challenging to
analyse and process data. For example, images, song lyrics, text documents (email, social media
post, video, etc.

3. Semi-structured data: It contains elements of both structured and unstructured data. It is not as
organised as structured data that precisely fits into the table, but it is easier to handle than
unstructured data. For example, HTML web pages, XML, and JSON files

11. What are the types of data used while developing an AI Project/Model?

1. Training data: Training data is the initial dataset used to train an AI model. It is a set of examples that
helps the AI model learn and identify patterns or perform particular tasks. We must ensure that the
data used to train the AI model is aligned with the problem statement and is sufficient, relevant,
accurate, and wide-ranging.

2. Testing data: Testing data is used to evaluate the performance of the AI model. It is the data that the
AI algorithm has not seen before. It allows for checking the accuracy of the AI model. Testing data
should represent the information that the AI model will encounter practically in real-world
situations.

12. What are Data Features?

Data features describe the type of information that will be collected in response to the problem
statement. In order to determine the data that is required, we first need to visualise the factors
influencing the problem statement. For this, we need to extract the data features for the problem
scope and find out the parameters that affect the problem statement directly or indirectly.

13. What is bioethics in the context of AI?

Bioethics in AI refers to ethical considerations when AI is used in biological or healthcare fields. It

ensures safety, privacy, and fairness while using AI in sensitive domains like medicine, genetics, or
diagnosis.

14. What is a framework? Give examples.

A framework is like a guide or a structure that helps organise complex ideas or processes into
manageable parts. It provides a step-by-step approach to solving problems or making decisions. For
example, a recipe is a framework for cooking a dish, or a school timetable is a framework for
organising classes.

In Artificial Intelligence (AI), frameworks guide the design, development, and use of AI systems to
make them reliable, effective, and safe.
 15. What is an Ethical Framework?

An ethical framework is a set of rules or principles that guide people to make fair, honest, and
responsible decisions. These principles ensure that actions are morally right and benefit everyone
involved, i.e., it helps us think about the impact of our decisions on individuals, society, and the
environment. In AI, ethical frameworks guide how AI systems are developed and used to avoid ha
ensure fairness, and respect human rights.

For example:

Doctors follow the principle of "do no harm" when treating patients.

Environmentalists promote sustainability to protect the planet for future generations.

B. Long answer type questions.

1. Explain the significance of the data exploration stage in the Al Project Cycle and how it
contributes to the success of Al projects.

The data exploration stage is crucial as it involves analysing and understanding the collected data to
discover patterns and insights. This process helps in several ways.

i. It simplifies complex data, making it easier to comprehend for both technical and non-technical
stakeholders.
ii. It uncovers hidden relationships or anomalies that might not be immediately apparent, providing
valuable insights.
iii. Data exploration aids in selecting suitable AI models for subsequent stages by identifying relevant
data features,
iv. It plays a vital role in communicating the findings to stakeholders effectively

2. Explain any five principles in the ethical framework of AI and their importance in ensuring
responsible AI development and usage.

The ethical framework of AI includes several key principles to ensure AI systems are developed and
used in a responsible, fair, and transparent manner. Five key principles are:

Transparency: All systems should be transparent, meaning their decision-making processes, data
usage, and underlying algorithms must be understandable to users. This helps build trust and allows
for accountability in AI decisions, ensuring that users can question and verify outcomes.

Fairness: Al must be designed to treat all individuals equitably, avoiding biases related to race,
gender, age, or other characteristics. Fairness ensures that AI systems do not perpetuate societal
inequalities, promoting equal opportunities for all.

Accountability: Developers and organisations must be responsible for the actions and outcomes of AI
systems. Accountability ensures that any harm caused by AI systems can be traced back to those
responsible, leading to corrective actions.

Privacy: Protecting user data is critical. All systems should handle personal information securely, with
proper consent and safeguards against misuse, ensuring that individuals' privacy rights are
respected.

Non-maleficence: This principle ensures that AI systems do not cause harm to individuals or society.
It requires developers to design AI with safety measures, preventing unintended consequences,
errors, or harmful biases from impacting users.
3. Discuss the different ethical frameworks applied in AI and their significance.

Various ethical frameworks guide the development of AI systems, each offering a unique perspective
on how AI should interact with society.

Deontological ethics focuses on following rules and duties, ensuring that AI systems respect laws and
principles, like privacy regulations.

Utilitarian ethics aims to maximise overall benefits, advocating for AI applications that serve the
greater good, such as in public health.

Virtue ethics focuses on the moral character of AI developers, encouraging qualities like fairness and
transparency in AI design.

Feminist ethics promotes inclusivity and equality, ensuring AI systems do not perpetuate biases.

4. Samarth attended a seminar on AI and has now been asked to write a report on his learnings
from the seminar. Being a non-technical person, he understood that the AI-enabled machine uses
data in different formats in many of the daily-based applications, but failed to sync it with the right
terminologies and express the details. Help Samarth define Artificial Intelligence by listing the
three domains of AI and the data used in each.

Data Science: This domain involves collecting, processing, and interpreting large datasets. An
application is using data analysis to create price comparison websites, helping users find the best
deals.

Computer Vision(CV): Computer vision enables machines to interpret visual information from
images and videos. An application of computer vision is its use in self-driving cars to detect and
analyse objects on the road.

Natural Language Processing (NLP): NLP focuses on understanding and processing human language.
Virtual assistants, like Siri, are an application of NLP, which is used to interpret voice commands and
provide responses.

5. What is the significance of the Al project cycle? Also, explain in detail about how data
acquisition is different from data exploration.

The AI Project Cycle is a structured framework that outlines the different stages of creating an AI
system. It is a systematic and step-by-step process for developing and implementing AI projects. The
different stages of the Al Project Cycle are:

i. Problem scoping
ii. Data acquisition
iii. Data exploration
iv. Modelling
v. Evaluation

Data Acquisition focuses on collecting the relevant data required for the AI system.

Data Exploration involves the exploration and analysis of the collected data to interpret patterns,
trends, and relationships.

6. Explain the importance of the 4Ws Problem Canvas in problem scoping.

The 4Ws Problem Canvas helps define the problem by identifying stakeholders, clarifying the
problem's nature, specifying its context, and highlighting the reasons for solving it. It provides a
structured framework for thorough problem analysis.
7. Discuss AI ethics and their importance.

AI ethics refers to a set of moral principles and techniques that form the guidelines for the
development and deployment of AI to ensure ethical and responsible use of the technology. As
humans, we have an inbuilt moral compass that tells us right from wrong. Unfortunately, artificial
intelligence lacks this ability. Al can only separate right from wrong based on the data that the
developer has fed with the labels 'right' and 'wrong'. If the developer has not considered the
outcome of each AI algorithm, it could lead to undesirable or harmful results.

8. Read the following paragraph carefully and categorise the highlighted terms under Data Science,
Machine Learning, Computer Vision, and NLP.

The latest technological advancements have made our lives convenient. Google Home, Alexa, and Siri
have been a huge help to non-tech-savvy people. Features like facial recognition and FaceLock have
added additional security to our gadgets. These advancements have also contributed to making our
needs more approachable and convenient. Now you can even check the prices with price
comparison websites and order groceries online with chatbots. Did you know that you can even find
out how you are going to look when you grow old? FaceApps and Snapchat filters have made this
possible."

 Computer Vision: Facial Recognition, FaceLock, and FaceApps

 Data Science: Price comparison websites

 NLP: Alexa, Siri, and Chatbots

 Machine Learning: Snapchat filters

9. Write any four applications of NLP.

The four applications of NLP are:

i. Email filters or spam filters: Gmail and other mailing platforms use email filters or spam
filters to separate unwanted or malicious messages, known as spam, from genuine emails.

ii. Virtual assistants and chatbots: We can use natural language to give instructions to the
virtual phone assistants, also called pocket assistants, like Siri or Google Assistant. Chatbots
use NLP to understand and respond to human language input in a natural way.

iii. Sentiment analysis: NLP is used to analyse and determine the sentiment or emotional tone
of text data, such as social media posts, customer reviews, and news articles. This helps
businesses gauge public opinion and customer satisfaction.

iv. Language translation: NLP powers machine translation systems like Google Translate. It
enables the automatic translation of text from one language to another, making cross-lingual
communication more accessible and efficient.

10. What is data visualisation? How does it help?

Data visualisation is the graphical representation of data and information using visual elements like
charts, graphs, or maps to make data easy to understand. Data visualisation helps us in the
following ways:
 i. It simplifies complex data, thus making it easier to comprehend.

ii. It helps gain a deeper understanding of the trends, relationships, and patterns present within
the data.

iii. It uncovers hidden relationships or anomalies (odd behaviour) that may not be immediately
apparent.

iv. It helps us select models for the subsequent AI Project Cycle stage.

v. It makes it easy to communicate insights to others, even non-technical people.

11. Explain various data visualisation techniques.

1. Bullet Graph
 One-line Description: A bullet graph is a variation of a bar chart with a primary bar layered
on top of a secondary bar to compare actual vs target values.
 Suitable for: Comparing performance against benchmarks or goals (e.g., sales, expenses,
performance indicators).

2. Histogram
 One-line Description: A histogram is a graphical display of data using bars to show the
frequency of values in specific numeric ranges.
 Suitable for: Continuous or interval data (e.g., height ranges, marks scored, customer age
distribution).

3. Scatter Plot
 One-line Description: A scatter plot displays data points for two variables on the X and Y axes
to show the relationship or correlation between them.
 Suitable for: Bivariate data (e.g., temperature vs ice cream sales, hours studied vs marks).

4. Tree Diagram
 One-line Description: A tree diagram represents hierarchical data using nodes and branches
to show relationships or classifications.
 Suitable for: Hierarchical or multi-level data (e.g., animal classification, organization chart).

5. Flowchart
 One-line Description: A flowchart is a diagram that shows the steps of a process or decision-
making using symbols and arrows in a logical sequence.
 Suitable for: Algorithms, workflows, processes (e.g., what to do if a device doesn’t work,
programming logic).

6. Pie Chart
 One-line Description: A pie chart is a circular chart divided into slices, where each slice
represents a category’s proportion of the total.
 Suitable for: Categorical data with proportions or percentages (e.g., favorite fruit, election
results).

7. Area Chart
 One-line Description: An area chart is like a line chart but with the area under the line filled,
emphasizing volume or magnitude over time.
 Suitable for: Time-series data showing trends and totals (e.g., company revenue over
months, rainfall data).
8. Bubble Map
 One-line Description: A bubble map shows data on a geographical map using bubbles, where
the size and position of each bubble represent values.
 Suitable for: Geo-spatial data (e.g., population by country, COVID cases by city, rainfall by
state).

9. Bar Graph
 One-line Description: A bar graph uses rectangular bars to show comparisons among
discrete categories or items.
 Suitable for: Categorical or discrete data (e.g., items sold in shops, number of students in
each section).

10. Venn Diagram

 One-line Description: A Venn diagram uses overlapping circles to show logical relationships
and shared elements between sets.
 Suitable for: Set-based and comparative data (e.g., students who like Math, Science, or
both).

11. Line Chart

 One-line Description: A line chart displays information as a series of data points connected
by straight lines to show trends over time.
 Suitable for: Time-series data or continuous data where the change in values is important
(e.g., temperature across days, stock prices).

12. Explain various sources of data.

Survey: A survey is a method of gathering specific information from a group of people by asking
them questions. They enable us to collect valuable data quickly and efficiently. Surveys can be
conducted on paper, through face-to-face or telephone interviews, or online forms. For example,
census surveys are conducted every year for population analysis,

Web Scraping: Web scraping is the process of collecting information from websites in an
automated manner. You can copy and paste the data from the website into a document and use it
as a data source. Web scraping can be done manually, but if a large number of websites need to be
accessed, some automated web scraping tools like BeautifulSoup or Scrapy can be used to help
speed up the process.

Sensors: Sensors are devices that detect and measure environmental conditions, such as
temperature, pressure, light, sound, and motion. They convert these physical parameters into
electrical signals or digital data that can be processed and analysed by AI systems. They are often
integrated with lot devices. For example, in a smart home, sensors can be used to measure the
room temperature, or in autonomous cars to detect objects in the surrounding area.

Cameras: Cameras such as cell phone cameras, webcams, CCTV, and surveillance systems can be
used to capture visual information in the form of images and videos. They play an important role in
the Computer Vision domain of AI by analysing and understanding visual data. For example,
capturing real-time videos of the surroundings through surveillance cameras.

Observations: It involves gathering data manually through direct observations of real-world events
as they happen. For example, counting the number of birds in an area in a particular season.

APIs: Application Programming Interfaces (APIs) are programs used by developers to acquire data
from other programs, services, or databases to extract relevant data required for the AI project.
Here, Data Acquisition is done automatically through special programs. For example, there is an AI
 project involving sentiment analysis, developers can use a social media API to access user posts or
comments on Facebook or X (Twitter).

13. How do bioethical principles apply when AI is used in fields like healthcare, life sciences,
genetic research/biotechnology, research, and environmental science, etc.?

Or, how are the core principles of Bioethics practiced across various fields?

Bioethics plays a critical role in many areas of healthcare, research, and biotechnology. It is
especially important when dealing with complex issues that involve potential risk to individuals,
communities, or society. Before are several key areas where bioethics is applied.

1. Medical ethics: Bioethics guides medical professionals in making ethical decisions in patient care.
For instance, doctors must navigate issues related to informed consent, end-of-life care, and the
use of experimental treatments. Medical bioethics helps ensure that patients' rights are respected
and that they are not subjected to unnecessary harm

2. Genetic research and biotechnology: Bioethics addresses the moral implications of genetic
modifications, cloning, and gene therapy. These technologies have the potential to cure genetic
diseases, but they also pose risks, including unintended consequences for future generations.

3. End-of-life decisions: Bioethics guides on end-of-life issues such as euthanasia, assisted suicide,
and the use of life-support technologies. These topics raise difficult questions about when to
continue treatment, when to stop, and how to ensure that individuals' wishes are respected while
minimising suffering.

4. Research ethics: Bioethics plays a central role in guiding the conduct of medical and biological
research. It ensures that experiments involving humans are conducted in a way that is ethical,
including obtaining informed consent and minimising risks to participants. The ethics of clinical
trials, animal research, and the use of new medical technologies are all areas of concern.

5. Environmental bioethics: Environmental bioethics deals with the moral obligations of humans to
the natural world. This branch of bioethics examines the ethical implications of environmental
practices such as resource extraction, pollution, and biotechnology in agriculture. It raises
questions about the sustainability of human activities and their impact on the environment.

14. What are the core principles of Bioethics?

Bioethics is guided by several fundamental principles that help professionals make ethical
decisions. These principles provide a framework for evaluating the consequences of medical
practices and biological research. They include:

1. Autonomy: This principle emphasises the right of individuals to make informed decisions about
their bodies and healthcare. Respecting autonomy means allowing patients to have control over
their medical choices as long as they are informed and capable of making decisions.

2. Beneficence: Beneficence refers to the ethical obligation to act in the best interests of others,
promoting their well-being and providing benefits. In healthcare, this principle requires medical
professionals to provide care that enhances health and alleviates suffering.

3. Non-maleficence: The principle of non-maleficence is based on the idea of "no harm". It calls for
healthcare providers to avoid causing harm to patients, whether through medical procedures,
negligence, or mistakes. It also requires careful consideration of the risks and benefits of treatments
and interventions.
 4. Justice: Justice in bioethics refers to fairness and the equitable distribution of healthcare
resources and services. It ensures that all individuals, regardless of their background, have equal
access to healthcare and treatment. The principle of justice also involves addressing disparities in
healthcare systems and ensuring that vulnerable populations are not exploited or neglected.

15. Define bioethics:

Bioethics is a field of study that addresses ethical issues arising from advances in biology, medicine,
and biotechnology. It involves the application of ethical principles to questions and dilemmas
related to life sciences, healthcare, and medical research. Bioethics aims to guide professionals,
policymakers, and society in making moral decisions when it comes to issues such as patient rights,
medical treatments, research ethics, and the use of new technologies in medicine.

16. How is bioethics multidisciplinary? How does it lay with other fields?

Bioethics is multidisciplinary, involving not only ethics but also law, social sciences. Philosophy and
theology. It seeks to balance scientific progress with moral responsibility. Ensuring that new
medical practices, discoveries, and technologies are used in ways that respect human dignity,
protect vulnerable populations, and promote justice and fairness.

17. Discuss the importance of an ethical framework in AI.

Need for an Ethical Framework of AI:

1. Prevent harm: Al can make mistakes or decisions that harm people. For example, a biased facial
recognition system could wrongly accuse someone of a crime. Ethical frameworks help prevent such harm
by ensuring fairness and accuracy.

2. Promote fairness: Al must treat everyone equally, regardless of their gender, race, or background. Ethical
guidelines ensure that Al does not discriminate against anyone.

3. Ensure transparency: Sometimes, AI systems make decisions that are hard to understand (e.g., why a loan
application was rejected). Ethical frameworks require developers to make AI decisions explainable and
transparent.

4. Respect privacy: Al often collects and uses personal data. Ethical frameworks ensure this data is handled
securely and responsibly, respecting people's privacy.

5. Build trust: People will only use Al if they trust it. Following ethical frameworks helps build trust by
showing that Al is used responsibly and for the greater good.

6. Address complex issues: As Al grows more advanced, it may face complex ethical dilemmas. For example,
in self-driving cars, should the car prioritise the safety of passengers or pedestrians? Ethical frameworks
help address such tough questions.

18. Explain how ethical principles contribute to the responsible development and deployment of
AI systems.
Or, what are the principles of the Ethical Framework in AI?

Principles of Ethical Frameworks in AI

 To ensure that AI systems are safe, fair, and trustworthy, ethical frameworks are guided by specific
principles. These principles address the potential risks and challenges of AI while maximising its
benefits for individuals and society.

1. Transparency: Transparency ensures that Al systems are open about how they function and how decisions
are made. Users should be able to understand the logic, data, and processes behind AI decisions.
Example: If an AI system denies a loan application, it should explain the reason (eg, low credit score, high
debt). Users should know what factors influenced the decision.

2. Fairness: Fairness in AI ensures that AI systems treat all individuals equally, without discrimination or bias.
It promotes equality by ensuring everyone receives the same treatment, regardless of factors such as race,
gender, age, or background. Fairness also helps prevent harm by minimising the risk of biased or
discriminatory outcomes in decision-making processes.
Example: In a recruitment process. Al should assess candidates based on qualifications and skills, not on
irrelevant factors like ethnicity or gender.

3. Inclusivity: Inclusivity ensures that AI systems are designed to serve diverse populations and consider the
needs of marginalised or underrepresented groups. It ensures that Al is accessible to all, including those
with disabilities or from different cultural backgrounds.
Example: accessible to people with disabilities, should assist in multiple languages, and be

4. Accountability: Inclusivity ensures that AI systems are designed to serve diverse populations and consider
the needs of marginalised or underrepresented groups. It ensures that Al is accessible to all, including
those with disabilities or from different cultural backgrounds.
Example: If an autonomous car causes an accident, accountability determines whether the manufacturer,
developer, or user is responsible and ensures that corrective actions are taken.

5. Privacy: Privacy ensures that AI systems protect user data from unauthorised access, misuse, or sharing. It
also means that data collection should be minimised and consent-based. It encourages users to engage
with AI systems without fear of losing control over their data.
Example: A healthcare AI system must securely store patient data and ensure it is used only for diagnosis
or treatment, not shared with advertisers or unauthorised parties.

6. Safety: Safety ensures that Al systems operate without causing harm to individuals, society, or the
environment. It includes testing, monitoring, and mitigating risks associated with AI use. It ensures Al does
not malfunction or cause unintended consequences.
Example: A robotic surgery system must be thoroughly tested to ensure it performs accurately, reducing
the risk of errors during operations.

7. Non-maleficence: Non-maleficence is the principle of "no harm". It ensures that AI systems are not used
to harm individuals or communities, whether intentionally or unintentionally. It ensures Al is used for good
purposes and not for harmful activities.
Example: AI should not be used to create deepfake videos that spread false information or damage
reputations.

8. Beneficence: Beneficence is the principle that Al should aim to do good and provide benefits to
individuals, communities, and society as a whole. It encourages the use of AI to solve societal challenges,
like improving healthcare or education.
Example: A systems that help predict natural disasters can save lives by providing timely warnings.

19.
What are the types of ethical frameworks in AI?

Ethical frameworks serve as guidelines to make moral decisions about the development and use of AI.
They help balance innovation with responsibility and ensure that AI technologies benefit humanity
without causing harm.
1. Deontological Ethics (Duty-based Ethics):
Deontological ethics focuses on the adherence to rules or duties regardless of the outcomes. In AI, this
framework emphasises following ethical principles and laws in the design and deployment of AI systems.

Example: Ensuring that Al respects user privacy and adheres to data protection laws, such as the General
Data Protection Regulation (GDPR) in Europe, is a form of applying deontological ethics. Even if it could be
more efficient to ignore privacy concerns for a better user experience, the duty to respect privacy takes
precedence.

2. Utilitarian Ethics (Outcome-based Ethics):

Utilitarianism suggests that the morally right action is the one that maximises overall happiness or well-
being. When applied to Al, this framework seeks to achieve the greatest benefit for the greatest number
of people.
Example: AI systems used in healthcare diagnostics aim to maximise the positive outcomes by diagnosing
diseases more accurately, potentially saving lives. However, there may be trade-offs, such as the risk of
errors or biases, which need to be carefully considered to ensure the overall benefit outweighs the
potential harm.

3. Virtue Ethics (Character-based Ethics):

Virtue ethics emphasises the character traits and moral integrity of the individuals developing or using AI.
It asks, "What kind of person should we be to make ethical decisions in AI?"
Example: In Agile design, developers should embody virtues like honesty, transparency, and fairness. If an
AI system is being used to make hiring decisions, it should be developed in a way that reflects virtues like
fairness and respect for all applicants, avoiding biased or discriminatory results.

4. Feminist Ethics:
Feminist ethics focuses on issues like equality, justice, and the importance of diverse voices in ethical
decision-making. It challenges the existing power structures and encourages an inclusive approach to
ethical AI development.
Example: AI systems should be designed with consideration for gender equality and should not reinforce
stereotypes or bias against marginalised groups. A feminist approach to AI ethics might advocate for more
diverse teams in AI development to ensure a broad range of perspectives are considered.

5. Care Ethics:
Care ethics is centred on the importance of empathy, relationships, and responsibility. In Al, care ethics
suggests that the technology should promote well-being, nurture relationships, and care for individuals'
needs.

20. Elaborate on the purpose and collaboration behind the development of IBM Watson for Oncology.

IBM Watson for Oncology was an AI system developed by IBM in collaboration with Memorial Sloan
Kettering Cancer Center. Its primary purpose was to assist oncologists by analyzing vast amounts of
medical data, including patient records, clinical trial results, and medical literature, to provide evidence-
based treatment recommendations for cancer patients. This helped in making informed decisions for
personalized cancer care.

21.
Explain the concept of 'Bias and Fairness' as an ethical issue in AI, specifically referencing the challenges
faced by IBM Watson for Oncology.

Bias and fairness in AI refer to the unintended discrimination or unfair outcomes that an AI system might
produce due to skewed or unrepresentative data it was trained on. In the context of IBM Watson for
Oncology, this was a concern because the system was trained on predominantly Western patient
 populations, which could lead to suboptimal or inappropriate treatment recommendations for patients
from diverse or minority populations, thereby perpetuating health disparities.

22.
Discuss the ethical challenge of 'Accountability' in AI systems. Why is it particularly difficult to address in
practice?
Accountability in AI refers to the challenge of assigning responsibility when an AI system makes an error or
provides an incorrect recommendation. It is difficult because AI's decision-making processes can be
opaque ('black boxes'), making it hard to trace how a particular conclusion was reached. This raises an
ethical dilemma: who should be held responsible – the developers, the data providers, the users, or the AI
itself?

23.
What do 'Transparency and Explainability' mean in the context of AI in medicine, and why are they
considered vital ethical considerations?

Transparency and explainability in AI refer to the ability to understand and interpret how an AI system
arrives at its conclusions or recommendations. In a medical setting, this is critically important because
clinicians need to trust and validate the AI's suggestions before applying them to patient care. Without
transparency, an AI's recommendation could be seen as a 'black box,' hindering a doctor's ability to
provide informed consent or explain treatment plans to patients effectively.

24.
Explain the ethical implications surrounding 'Privacy and Data Use' when AI systems handle sensitive
patient information. Provide examples of the type of data involved.

Privacy and data use are paramount ethical concerns in AI, especially when dealing with sensitive
information like medical records, test results, and genetic data. The vast amounts of patient data used by
AI systems like Watson necessitate strict adherence to data protection standards (e.g., HIPAA). Key issues
include ensuring patient consent for data use, maintaining data security, preventing unauthorized access,
and avoiding misuse of personal health information.

25.
Define the AI project cycle and explain its Various stages.

The AI Project Cycle is a structured framework that outlines the different stages of creating an AI system. It
is a systematic and step-by-step process for developing and implementing AI projects. Each stage involves
careful planning and execution to design and deploy AI solutions.

The different stages of the Al Project Cycle :

1. Problem scoping: The first step is to understand and define the problem that we want Al to solve. Problem
scoping is the stage where we set clear goals and outline the objectives of the Al project. It includes
precisely outlining the issues, defining them explicitly. Identifying their causes and developing a plan to fix
them.

2. Data acquisition: This stage focuses on collecting the relevant data required for the algorithm. Since this
data forms the base of your project, care must be taken to ensure that the data is collected from reliable
and authentic sources.

3. Data exploration: This stage involves the exploration and analysis of the collected data to interpret
patterns, trends, and relationships. The data is in large quantities. In order to easily understand the
patterns, you can use different visual representations, such as graphs, databases, flowcharts, and maps.

4. Modelling: After exploring the patterns, you need to select the appropriate AI model to achieve the goal.
This model should be able to learn from the data and make predictions.
5. Evaluation: The selected Al model now needs to be tested, and the results need to be compared with the
expected outcome. This helps in evaluating the accuracy and reliability of the model and improving it.

6. Deployment: Once the evaluation is complete and generates accurate results, the AI model can be
deployed in the real world.

26.
Suhana works for a company where she was assigned the task of developing a project using the Agile
project cycle. She knew that the first stage was scoping the problem. Help her list the remaining stages
that she must go through to develop the project.

Since Suhana has already completed Problem Scoping, the remaining stages are:
1. Data Acquisition – Collecting data relevant to the problem.
2. Data Exploration – Understanding and analyzing the data.
3. Modelling – Building AI models using data.
4. Evaluation – Testing and validating the model’s accuracy.
5. Deployment – sending the model to the market.

27.
What is the 4Ws Problem Canvas? Explain its parameters.

The 4Ws Problem Canvas is a structured framework that helps in problem scoping. It helps in identifying
the key elements related to the problem. It provides a structured framework for analysing and
understanding a problem. It consists of four crucial parameters that we need to know while solving a
problem: Who? What? Where? Why?

Who? :
The "Who?" block helps identify the people who are directly or indirectly affected by the problem. These
people are referred to as stakeholders. Stakeholders are involved in the problem and will benefit from
the solution we arrive at.

What? :
The "What?" block helps determine if the problem exists. It helps to clearly define this newspaper, social
media posts, reports, etc.

Where? :
The "Where?" block helps identify the situations where the problem is observed or has an impact.
These could include physical locations like cities or specific areas, online platforms, contexts such as
educational institutes or public spaces, or whether it has a global or local impact.

Why? :

In the "Why?" canvas, think about why the given problem is worth solving and how the solution would
benefit stakeholders as well as society. This could include a better quality of life, an increase in efficiency,
cost savings, enhanced safety, long-term sustainability, or other positive outcomes.

28. What is a problem statement template?

The problem statement template aids in summarising all the key points mentioned in the 4Ws canvas
into a single template, which enables further analysis and decision making. This template makes it
simple to understand and remember the important aspects of the problem.

Who?
 Who is having a problem/stakeholder?

What do you know about them?

What?
What is the problem / and its nature?

How do you know that it is a problem? Is there any evidence?

Where?
What is the context?

What is the location where the stakeholders experience this problem?

Why?
Why is this problem worth solving and of value to stakeholders?

How will the situation of stakeholders change?

Case Study 1: AI-Based Waste Segregation System

Case Study:
At EcoSmart School, waste segregation has been a big challenge. Students often throw biodegradable
and non-biodegradable waste into the same bins, making recycling difficult. The school’s eco-club
suggested using AI to solve this problem. They plan to install smart dustbins with cameras and an AI
system that identifies the type of waste and guides students to dispose of it correctly. The system will
also collect data on daily waste habits to encourage better environmental practices.
4W Canvas Answers:
 Who: Students, Eco-Club Members, and School Management

 What: Students throw bio/non-biodegradable waste material, no clear waste segregation

process in the school. We see waste material lying in the school even after cleaning.
 Where: In school canteens, corridors, classrooms, and common areas
 Why: To ensure proper waste segregation and promote environmental responsibility among
students

Case Study 2: Clean Campus Initiative

Context:
The school faces issues with littering, even though there are dustbins. The management wants a system
that detects littering behavior and alerts staff.
Task for Students:
Draw the 4W Canvas to explore how AI-based CCTV image processing can help.
 Who: School authorities, Students
 What: Littering is happening in school, students are not throwing waste in the dustbin, waste is
seen in the classroom and corridors
 Where: School playgrounds, corridors, canteen
 Why: To promote cleanliness and responsible behavior, Monitoring and detecting littering
behavior in students
Case Study 3: AI-Based Career Guidance
Context:
Class 10 students feel confused about which stream to choose for Class 11. The school wants to build a
chatbot that asks questions and suggests suitable streams based on interest and performance.
Task for Students:
Draw the 4W Canvas for this proposed AI career bot.
 Who: Class 10 Students, School Counselors
 What: Students said that they are not able to decide which stream to opt for in class 11
 Where: In schools where Class 11 students are experiencing this problem
 Why: To help students make informed stream choices

Case Study 4: Library Book Finder

Context:
Students often waste time looking for books in the library. The school librarian wants to create an AI
system that helps students find books using voice or text input.
Task for Students:
Use the 4W Canvas to frame the problem clearly.
 Who: Students, Librarian
 What: Students are not able to find the books in the library quickly. The average period is wasted
on finding the right book
 Where: School Library
 Why: To save time and improve library usage

Case Study 5: Healthy Canteen Choices

Context:
The school wants to promote healthy eating. They plan to use an AI system to suggest healthy
alternatives when students order junk food in the digital canteen app.
Task for Students:
Create a 4W Canvas to understand how this system can help students make better choices.
 Who: Students, Canteen Manager
 What: Students order junk food in the canteen, and the order rate for junk food is high
 Where: School Canteen (via digital menu)
 Why: To encourage nutritious eating habits and stay healthy

Case Study 6: AI-Based Homework Tracker

Case Study:
At Harmony Public School, many students forget to complete or submit their homework on time,
especially when they have multiple assignments. Teachers find it hard to track pending work for each
student. The school wants to introduce an AI-powered homework tracker that sends automated
reminders to students and helps teachers view submission patterns. The tool can also analyze which
subjects students struggle with based on late submissions and alert teachers to intervene early.
4W Canvas Answers:
 Who: Students and Teachers
 What: Many students forget to complete or submit their homework on time, especially when
they have multiple assignments. Teachers find it hard to track pending work for each student.
 Where: In the school, the teacher needs to give multiple reminders to the students
 Why: To help students manage their homework better and reduce late submissions, while
allowing teachers to give timely support
Case Study 7: AI-Based Lost & Found System
Case Study:
Students at Blue Bells School frequently misplace their belongings like bottles, sweaters, and
lunchboxes. Managing the Lost & Found desk has become a challenge. The school plans to install an AI-
based image recognition system that matches lost items with student profiles to automate notifications.
4W Canvas Answers:
 Who: Students, Teachers, and Lost & Found Staff
 What: Student very often forget their belongings like bottles, sweaters, ties, etc, which need to
be stored and remain unclaimed
 Where: School’s Lost & Found area
 Why: To reduce the time and effort spent on manually managing lost items

Case Study 8: Classroom Mood Monitor

Case Study:
Teachers at Crestview School struggle to identify students who are emotionally unwell or disengaged in
class. The school is testing an AI tool that detects students’ emotional states using facial expressions,
helping teachers and counselors take timely action.
4W Canvas Answers:
 Who: Students, Teachers, and School Counselors
 What: Teachers face difficulty in identifying students who are emotionally unwell or disengaged
in class
 Where: Inside classrooms through existing cameras or manual surveillance/supervision
 Why: To support students’ emotional well-being and provide early help to the teacher

Case Study 9: AI-Enabled Library Book Return Reminder

Case Study:
At New Horizon School, many students borrow library books but forget to return them on time. This
causes delays for other students who wish to read those books. The librarian maintains a manual record
and sends reminders occasionally, but it’s not always effective. To solve this, the school wants to develop
an AI system that tracks book due dates, analyzes student borrowing habits, and sends personalized
reminders. For instance, if a student often delays returns, the AI can send earlier or more frequent
reminders. Over time, the system learns patterns and suggests books based on the student’s interests
and reading history, too.
4W Canvas:
 Who: Students and the school librarian
 What: Many students borrow library books but forget to return them on time. This causes delays
for other students who wish to read those books
 Where: In the school library system and student dashboard
 Why: To improve timely book returns and optimize library usage for all students

Case Study 10: AI-Powered School Bus Delay Alert System

Case Study:
Parents of students at Mount View School are often concerned when the school bus is delayed due to
traffic or weather. Currently, there is no system to alert them in real-time. The school plans to implement
an AI-powered system that uses GPS and traffic data to predict delays and send automatic notifications
to parents and school staff. It will also learn from previous delay patterns and weather forecasts to
improve future predictions. This system aims to enhance safety, communication, and parent trust.
4W Canvas:
 Who: Students, Parents, and the School Transport Department
  What: The School bus is often delayed due to traffic or weather. Currently, there is no system to
alert them about the delay in the arrival of the bus.
 Where: Travelling to school and returning
 Why: To ensure safety, reduce anxiety, and improve transport communication

Case Study 11: AI for Student Noise Monitoring in Classrooms

Case Study:
In Oak Tree International School, teachers face challenges in managing noise levels in classrooms during
group activities. Some groups become too loud, disturbing others. The school plans to introduce an AI
tool that uses sound sensors to monitor classroom noise in real time. If a group exceeds acceptable
levels, the system sends a gentle visual cue on the smartboard (like a traffic light turning orange or red).
Teachers can later review noise data and patterns to adjust group seating or give feedback. This
encourages self-discipline, especially in lower grades, while also supporting classroom management.
4W Canvas:
 Who: Students and Teachers
 What: teachers face challenges in managing noise levels in classrooms during group activities.
Some groups become too loud, disturbing others.
 Where: Inside classrooms using smartboard systems and sound sensors
 Why: To maintain a positive learning environment and encourage responsible behavior during
group activities

29. Write a note on evaluation.

Evaluation is the process of assessing the reliability and efficiency of any AI model based on the output
generated by feeding the model with the testing data and comparing it with the expected outcome.
Once a model has been developed and trained, the model is tested using a dataset called testing data
the efficiency and effectiveness of the model are determined on the basis of various parameters that
help assess its overall performance. Evaluation happens at various stages of the Al Project Cycle to
identify areas for improvement and make necessary adjustments to achieve the project goals.
The parameters are:
 Accuracy
 Precision
 Recall
 F1 Score

30. Write a note on deployment.

Al project deployment is the process of implementing an AI model in a real-world scenario. The model is
integrated into the desired software or system and packaged in such a way that it can be used for
practical applications. The goal is to make the AI model useful for solving real-world problems.

31. Explain the role of Computer Vision in Al. How does Computer Vision enable machines to process
visual information, and what are its applications in areas like self-driving cars and facial recognition?

32. Explain the domains of Al and describe their contributions to various industries with relevant
examples.

Artificial intelligence (AI) uses training data to develop or improve the intelligence of Al model or
machine. Depending on the applications for which the AI algorithm is developed, different dataset types
are given to Al models during their training, with respect to the kind of data fed into the Al model. Al
models can be broadly categorized into three domains

 Data Science

 Computer Vision (CV)

  Natural Language Processing (NLP)

AI coordinates with these domains to replicate human intelligence.

1. Data Science

Data Science refers to a multidisciplinary field that includes the collection, processing, analysis, and
interpretation of large and varied datasets to extract meaningful information and facilitate decision-
making for Al applications.

Big Data refers to the storage and processing of massive sets of data. The larger the amount of data the
AI systems can access, the more they can learn. Therefore, they can produce more accurate results. Big
data is the fuel for artificial intelligence. The objective of this domain is to collect a large amount of data,
maintain datasets, and derive meaning from them. The information can be used to make decisions.

Applications of Data Science

The following are some common applications of Data Science:

i. Price comparison websites: Websites, like Compare India, Gadgets 360, etc., use big data to provide a
comparison of the prices of products from multiple vendors at one place.

ii. Search engines: Search engines, like Google, Bing, etc., collect massive amount of data from various
sources, including search queries, web pages, and user behaviour, and analyse this data to provide
better search results to users.

iii. Website recommendations: Companies, like Amazon, Flipkart, etc., use large volumes of data to
recommend products based on previous search results for a user. They display products and
advertisements based on user data to enhance user engagement and experience.

2. Computer Vision

Computer Vision (CV) domain of AI enables computers to perceive and extract meaningful insights from
digital content (Uke photos and videos), analyse the information, and then/make decisions from it the
same way as humans go. This multidisciplinary domain's primary goal is to use sensors, computers, and
machine learning methods to replicate and automate various aspects of human vision.

Applications of Computer Vision

The following are some common applications of Computer Vision:

i. Self-driving automobiles: Self-driving automobiles use Computer Vision extensively. Automated cars
from/companies like Tesla can detect the 360-degree movements of pedestrians, cyclists, vehicles,
etc. Computer Vision/helps them detect and analyse objects in real-time and/take decisions like
breaking, stopping, or keep driving,

ii. Facial recognition: Facial recognition is also an application of Computer Vision. Facial recognition is a
technology that is capable of identifying a person from a digital image or video. It is used as security
for unlocking devices like cell phones and tablets, as well as by investigation agencies to/identify
criminals.

3. Natural Language Processing

Natural Language Processing (NLP) is the ability of computers to understand and interpret human
language and text, i.e., natural language. The goal of NLP is to take a verbal or written input, process it,
and understand it.
 Applications of Natural Language Processing

The following are some common applications of NLP

i. Email filters or spam filters: Gmail and other mailing platforms use email filters or spam filters to
separate unwanted or malicious messages, known as spam, from genuine emails.

ii. Sentiment analysis: NLP is used to analyse and determine the sentiment or emotional tone of text
data, such as social media posts, customer reviews, and news articles. This helps businesses gauge
public opinion and customer satisfaction.

iii. Virtual assistants and chatbots: We can use natural language to give Instructions to the virtual phone
assistants, also called pocket assistants, like Siri or Google Assistant. Chatbots use NLP to understand
and respond to human language input in a natural way.

iv. Text summarisation: NLP can automatically summarise long pieces of text by extracting key
information and condensing it into shorter versions. This is valuable for tasks, like generating concise
news summaries or creating executive summaries of lengthy documents.

v. Language translation: NLP powers machine translation systems, like Google Translate. It enables the
automatic translation of text from one language to another, making cross-lingual communication more
accessible and efficient.