ARTIFICIAL

INTELLIGENCE

11/17/2024 SA-MIS-3-8 1
What is Artificial Intelligence?
 AI refers to computer systems/machines capable of
performing complex tasks that historically only a
human could do, such as reasoning, making
decisions, or solving problems.
 It is the effort to develop computer-based systems
that can behave as humans
 It is the science and engineering of making
intelligent machines
 Simulation of human intelligence by machines

11/17/2024 SA-MIS-3-8 2
What is AI? (Cont’d)
 Basic idea
The computer can be programmed to perform some
of the same logical reasoning tasks as a human
 The term AI was coined in the 1950s
 AI was founded as an academic discipline in 1956

11/17/2024 SA-MIS-3-8 3
What is Human Intelligence (HI)?
 HI is the intellectual capability of humans,
which is marked by complex cognitive feats
and high levels of motivation and self-
awareness.
Humans are able to learn, form concepts,
understand, and apply logic and reason.
 HI is a capability to recognize patterns, plan,
innovate, solve problems, make decisions,
and use language to communicate.

11/17/2024 SA-MIS-3-8 4
Some HI Capabilities
HI is a way of reasoning
 HI can be described as the
application of rules based on human
experience and genetics
 Rules are part of the knowledge
carried by humans
 If x, then y

11/17/2024 SA-MIS-3-8 5
HI Capabilities (Cont’d)
HI is a way of behaving
Behaving in the realm of cultural
and social restrictions (values)

11/17/2024 SA-MIS-3-8 6
HI Capabilities (Cont’d)
HI includes the development and
use of metaphors and analogies
 Ability to develop associations
 Ability to use metaphors and
analogies such as “like” and “as”
 “Common sense” or “generality”

11/17/2024 SA-MIS-3-8 7
HI Capabilities (Cont’d)
HI includes the creation and use of
concepts
 Ability to impose conceptual apparatus on
the world
 E.g. Cause and effect and time

 AI refers to an effort to develop

machines that can learn, understand
reason, behave, compare, conceptualize,
and use human language.
11/17/2024 SA-MIS-3-8 8
AI Applications
 AI Covers an ever-changing set of
capabilities as new technologies are
developed including machine learning and
deep learning
 AI is widely used throughout industry,
government, and science

11/17/2024 SA-MIS-3-8 9
AI Applications (Cont’d)
Advanced web search engines
 Google Search
Recommendation systems
 YouTube
 Amazon
 Netflix (online movies)
Understanding human speech
 Google Assistant
 Siri (Apple)
 Alexa (Amazon)
11/17/2024 SA-MIS-3-8 10
AI Applications (Cont’d)
 Self-driving cars
 Waymo (Alphabet)
 Tesla
 Generative and creative tools
 ChatGPT (OpenAI)
 AI art
 Spelling/grammar suggestions
 MS Office
 Grammarly
 Facial recognition technology
 BioID
 Paravision
11/17/2024 SA-MIS-3-8 11
 Major AI Techniques/Tools
 Expert systems
 Machine learning
 Neural networks and deep learning
 Genetic algorithms
 Natural language processing
 Computer vision systems
 Robotics
 Intelligent agents
Source: Laudon, K.; Laudon, J.: Management Information Systems, 2019

11/17/2024 SA-MIS-3-8 12
Expert Systems
 Represent the knowledge of experts as a set
of rules and facts
 A knowledge-based system that uses its
knowledge about a specific application area
to act as an expert consultant to end users
 The advice can be to operational process or
decision-making process
 Have some reasoning capabilities

11/17/2024 SA-MIS-3-8 13
Expert Systems (Cont’d)
Using Expert Systems
 Involves an interactive computer-based session
 Solution to a problem is explored with the ES
acting as a consultant to an end user
 User asks questions
 ES searches its knowledge base for facts and
rules
 Explains its reasoning process
 Gives expert advice

11/17/2024 SA-MIS-3-8 14
Machine Learning (ML)
ML is the ability to learn from data and
past experiences to identify patterns and
make predictions with minimal human
intervention.
ML uses data and algorithms to imitate the
way that humans learn
 Can identify patterns in very large
databases without explicit programming

11/17/2024 SA-MIS-3-8 15
ML Approaches
 Supervised Learning
 Requires a human to label the datasets
 The computer is presented with example inputs
and their desired outputs, given by a "teacher“
(human), and the goal is to learn a general rule
that maps inputs to outputs
 Unsupervised Learning
 No labels are given to the learning algorithm
 Analyzes a stream of data and finds patterns and
makes predictions without any other guidance
11/17/2024 SA-MIS-3-8 16
ML Approaches (Cont’d)
 Reinforcement Learning
 The program interacts with a dynamic
environment in which it must perform a
certain goal
Example:
 Driving a vehicle
 Playing a game against an opponent

 The program is provided feedback that's

analogous to rewards, which it tries to
maximize

11/17/2024 SA-MIS-3-8 17
Neural Networks
 NN are Neuro-computer systems whose
architecture is based on the human brain’s
neuron structure
 Attempts to emulate the processing
patterns of the biological brain
 Simulates the way neurons interact to
process data and learn from the
experience

11/17/2024 SA-MIS-3-8 18
Deep Learning (DL)
 DL uses multiple layers of neural networks
to simulate the behavior of the human brain
 Can recognize complex patterns in pictures, text,
sounds

11/17/2024 SA-MIS-3-8 19
Natural Language Processing (NLP)
 Systems that can understand and analyze
natural human language
 Being able to talk to computers in
conversational human language and have
“understand” us as easily as we understand
each other
 Involves research and development in
 Linguistics
 Psychology
 Computer science and other disciplines

11/17/2024 SA-MIS-3-8 20
Computer Vision Systems
 Systems that can view and extract
information from real-world images
 Can derive meaningful information from
digital images, videos and other visual
inputs
Applications:
 Facial recognition
 Self-driving cars

 Medical anomaly detection

11/17/2024 SA-MIS-3-8 21
Robotics
 Produces robot machines with computer-
controlled humanlike physical capabilities
 Designed to give robots the power of
 Sight or visual perception
 Touch or tactile capabilities
 Dexterity of skill in handling and manipulation
 Locomotion or the physical ability to move over
any terrain
 Navigation or the ability to properly find one’s
way to a destination

11/17/2024 SA-MIS-3-8 22
GPT
 GPT (Generative Pre-trained Transformers)
are Neural network-based language
prediction models
 Based on the semantic relationships between
words in sentences
 Text-based GPT models are pre-trained on a
large corpus of text which can be from the
Internet
 The pre-training consists in predicting the next
word or punctuation

11/17/2024 SA-MIS-3-8 23
GPT (Cont’d)
 GPT models accumulate knowledge about the
world, and can then generate human-like text
by repeatedly predicting the next word
 A subsequent training phase makes the model
more truthful, useful, and harmless
 Current GPT models are still prone to
generating falsehoods called "hallucinations"

11/17/2024 SA-MIS-3-8 24
GPT (Cont’d)
 Current GPT models include:
 ChatGPT (OpenAI)

 Bard (Google)

 Copilot (Microsoft)

 Claude (Anthropic)

 Multimodal GPT can process different types of

data such as images, videos, sound and text
 GPT-4 (OpenAI)

11/17/2024 SA-MIS-3-8 25
Types of Artificial Intelligence (AI)
AI Based on Capabilities
 Weak AI or Narrow AI
 General AI
 Super AI
AI Based on Functionality
 Reactive Machines
 Limited Memory
 Theory of Mind
 Self-Awareness
11/17/2024 SA-MIS-3-8 26
AI Based on Capabilities
 Weak AI or Narrow AI
 Able to perform a dedicated task with intelligence
 Currently available AI
 Cannot perform beyond its field or limitations, as it is
only trained for one specific task
 Some Examples:
 Playing chess (IBM Computer defeated the world
chess champion Garry Kasparov in 1997)
 Purchasing suggestions on e-commerce site
 Self-driving cars
 Speech recognition
 Image recognition
11/17/2024 SA-MIS-3-8 27
AI Based on Capabilities (Cont’d)
 General AI
 Could perform any intellectual task with
efficiency like a human
 Currently, there is no such system exist
 Still under research

11/17/2024 SA-MIS-3-8 28
AI Based on Capabilities (Cont’d)
 Super AI
 A level of Intelligence of Systems at which
machines could surpass human intelligence
 Can perform any task better than human
 Some capabilities include the ability to think, to
reason, solve puzzles, make judgments, plan, learn,
and communicate by its own
 Still a hypothetical concept of Artificial Intelligence

11/17/2024 SA-MIS-3-8 29
AI Based on Functionality
 Reactive Machines
 Purely reactive machines
 Do not store memories or past experiences for
future actions
 Only focus on current scenarios and react on it
as per possible best action
 Examples:
 Spam filters
 Recommendations on streaming services
 Google’s AlphaGo game player

11/17/2024 SA-MIS-3-8 30
AI Based on Functionality (Cont’d)
 Limited Memory
 Can store past experiences or data
 Can use stored data for a limited time period
 Examples:
 Self-driving cars
• Can store recent speed of nearby cars, the distance of
other cars, speed limit, and other information to
navigate the road
 Chatbots
• Use natural language processing to create
humanlike conversational dialogue, can store
past conversations to respond to requests
11/17/2024 SA-MIS-3-8 31
AI Based on Functionality (Cont’d)
 Theory of Mind
 Theory of mind refers to the capacity to understand
other people by ascribing mental states to them
 Knowledge that others' beliefs, desires, intentions,
emotions, and thoughts may be different from one's own
 Computers should understand the human emotions,
intentions, beliefs, and be able to interact socially like
humans
 Still not developed
 Lots of efforts and improvements are made for
developing such AI machines

11/17/2024 SA-MIS-3-8 32
AI Based on Functionality (Cont’d)
 Self-Awareness
 These machines will have their own
consciousness, sentiments, and self-awareness
 Super intelligent
 Smarter than human mind
 Still a hypothetical concept

11/17/2024 SA-MIS-3-8 33