CHAPTER 1 - INTRODUCTION
Slide 1 - What is AI
● Artificial Intelligence (AI) is the science and engineering of making machines
intelligent, especially computer programs that can think, learn, and make decisions
like humans.
● AI is developed by studying how the human brain works—how people learn, solve
problems, remember, and reason—and using that understanding to build computer
systems that can do similar tasks.
● AI is the science of making machines think and act like humans. It’s about teaching
computers to make decisions, learn from experience, and solve problems.
Main Goal:
To understand and build intelligent systems.
Definitions of AI (4 Types)
An electronic machine or computer thinks rationally. If the system thinks the right thing, the
system is rational.
Category Description Simple Example
Think like humans Tries to mimic how humans think A robot that solves puzzles like
people
Think rationally Tries to think logically and correctly A GPS choosing the fastest route
Act like humans Tries to behave like people A chatbot that talks like a person
Act rationally Takes best actions to reach a goal Self-driving car avoiding traffic
Slide 2 - Why AI is imp
● It helps us understand how intelligence works.
● It can perform useful tasks like diagnosing diseases or driving cars.
● It’s used in many fields: computer science, psychology, mathematics, linguistics
(language), and neuroscience (the study of the brain).
Examples of AI Applications:
● Playing chess (like Deep Blue)
● Writing poems
● Recognizing speech
● Recommending videos (like YouTube or Netflix)
● Diagnosing patients in hospitals
Slide 3 - Foundation of AI
Artificial Intelligence (AI) is not just built from computer science alone—it is based on knowledge
and techniques from many other fields. These include:
● Mathematics – for logic, probability, and decision-making
● Philosophy – for understanding how the mind works and how humans think
● Psychology – for learning how humans and animals process information
● Linguistics – for understanding how language works
● Computer Science and Engineering – for building machines and writing programs
These fields help us understand how intelligent behavior works, and then use this
understanding to build intelligent machines.
Explanation:
● Mathematics helps with logic and making smart decisions.
● Philosophy asks deep questions about thinking and intelligence.
● Psychology helps us understand how humans and animals learn.
● Linguistics explains how language works. And of course,
● Computer science and engineering help us build smart machines and write programs.
Together, these fields make AI possible.
Slide 4 - Turing Test
The Turing Test is a method to check whether a computer can behave like a human.
Invented by Alan Turing in 1950.
How it works:
● A human talks (by typing) to both a machine and another human.
● If the human cannot tell which is which, the machine is said to have passed the test.
To pass the Turing Test, a computer needs:
● Natural Language Processing – to understand and speak English (or any language)
● Knowledge Representation – to remember facts and information
● Reasoning and Problem Solving – to think and draw conclusions
● Machine Learning – to learn from past experiences
● Computer Vision (for full test) – to "see" the world
● Robotics (for full test) – to move and interact physically
Turing Test Examples
1. Eugene Goostman (2014)
● Chatbot posing as a 13-year-old Ukrainian boy.
● Convinced 33% of judges it was human in a test by the University of Reading.
● Claimed as one of the first to “pass” the Turing Test (though controversial).
2. ELIZA (1960s)
● One of the earliest chatbots, made by Joseph Weizenbaum.
● Simulated a psychotherapist by rephrasing user inputs.
● Some users believed it was truly human, though it used simple pattern matching.
3. PARRY (1972)
● Simulated a patient with paranoid schizophrenia.
● More advanced than ELIZA, and psychiatrists couldn’t always tell it was a machine.
4. Cleverbot (2008–present)
● Learns from human conversations online.
● In some tests, judged human by over 50% of people.
● Conversations can be surprisingly human-like at times.
5. ChatGPT (2020s)
● Advanced AI developed by OpenAI.
● While not designed specifically to pass the Turing Test, it often fools users in casual
conversation.
● Shows strong abilities in natural language, knowledge, and reasoning.
While some AI programs have claimed to pass the test in specific settings, none have fully
met the original idea proposed by Alan Turing
1. Eugene Goostman (2014)
● A chatbot simulating a 13-year-old Ukrainian boy
● Claimed to pass the test by fooling 33% of judges in a specific event
● But critics say it used tricks (language barrier, young age) to avoid hard questions
● Not considered a "true" pass by AI experts
2. Chatbots like ChatGPT or Google Gemini
● Can sometimes fool people in casual conversation
● But still make factual errors, lack human-level reasoning, and show robotic
patterns
WEAK & STRONG AI
Type Description Example
Narrow AI Limited to specific tasks Siri, Google Maps
(Weak AI)
(Strong AI) Can think, feel, reason like a human (theory No real-world example yet
only)
Slide 5 - History of AI
The Gestation Period (1943–1955)
● 1943: First model of artificial neurons (McCulloch & Pitts)
● 1950: Turing Test proposed (Alan Turing)
● 1956: “AI” term introduced at Dartmouth Workshop
Early Enthusiasm (1952–1969)
● 1957: General Problem Solver (GPS) – solved human-like problems
● 1958: LISP language – became the main AI programming tool
● 1959: Geometry Theorem Prover – solved math problems
● 1968: ANALOGY program – solved visual IQ test questions
A Dose of Reality (1966–1973)
● AI struggled with real-world problems
● Understanding language was harder than expected
● Optimism faded AI Winter (less funding, slower progress)
Knowledge-Based Systems (1969–1979)
● Example: DENDRAL – helped chemists identify molecules
● AI used for specific fields like medicine and engineering
Slide 6 - History of AI
AI Becomes an Industry (1980–Present)
● 1981: Japan launched Fifth Generation Computer Project
● AI used in:
○ Business automation
○ Robotics
○ Medical diagnosis
Return of Neural Networks (1986–Present)
● Neural networks became popular again
● Used in voice & image recognition
● Key scientists: Hinton & Rumelhart
Recent Developments (2000s–Present)
● Technologies like:
○ Speech & handwriting recognition (HMMs)
○ Smart decision-making (Bayesian networks)
○ Machine Learning & Deep Learning
○ Used in chatbots, self-driving cars, Netflix, YouTube, and more!
HISTORY OF Artificial Intelligence (AI)
1.5.1 The Gestation Period (1943–1955)
This is the "birth phase" of AI, when foundational theories were first developed.
Key Events:
● 1943: Warren McCulloch and Walter Pitts created the first model of artificial neurons.
○ Combined neuroscience + propositional logic + computation theory.
○ They showed how a simple network of neurons could perform logical reasoning.
● 1950: Alan Turing proposed the Turing Test, which became a core idea in AI (can
machines think?).
● 1956: John McCarthy, along with Marvin Minsky, Claude Shannon, and Nathaniel
Rochester, organized the Dartmouth Workshop.
○ First time the term "Artificial Intelligence" was officially used.
○ This workshop is considered the starting point of AI as a field.
1.5.2 Early Enthusiasm & Great Expectations (1952–1969)
This was a period of rapid progress and excitement in AI development.
Major Developments:
● 1957: General Problem Solver (GPS) by Herbert Simon & Allen Newell
○ Designed to solve a wide range of problems like humans do.
● 1959: Geometry Theorem Prover by Herbert Gelernter
○ Solved complex geometry problems, proving AI could handle mathematics.
● 1958: LISP programming language invented by John McCarthy
○ Became the main programming language for AI for many decades.
● 1968: ANALOGY Program by Tom Evans
○ Solved visual IQ test problems (figural analogies).
1.5.3 A Dose of Reality (1966–1973)
Despite early success, AI systems started facing limitations.
Challenges:
● Systems couldn’t scale well to real-world problems.
● Language understanding and general intelligence proved harder than expected.
● Famous quote by Herbert Simon (1957):
“There are now machines that think, learn, and create.”
But such optimism turned out to be premature.
Result:
AI progress slowed down, leading to the first “AI winter” – a time of reduced funding and
interest.
1.5.4 Knowledge-Based Systems (1969–1979)
AI research turned to domain-specific expert systems – programs that simulate human
experts in narrow fields.
Notable Example:
● DENDRAL Project:
○ Helped chemists identify unknown molecules by analyzing mass spectrometry
data.
○ One of the first expert systems in AI history.
● Other expert systems were created for medicine, engineering, and more.
1.5.5 AI Becomes an Industry (1980–Present)
AI shifted from a research topic to real-world commercial use.
Key Events:
● 1981: Japan launched the Fifth Generation Computer Systems Project.
○ Aimed to develop machines with reasoning and logic using Prolog.
● AI used in healthcare, banking, and industry.
Impact:
AI funding and investment exploded, especially in the U.S. and Japan.
1.5.6 The Return of Neural Networks (1986–Present)
Neural networks made a comeback, leading to what we now call deep learning.
Contributors:
● Inspired by how the human brain works (biological neurons).
● Can handle large datasets and recognize patterns—used in voice and image
recognition.
● Key scientists: Hinton & Rumelhart
1.5.7 Recent Developments (2000s – Present)
AI has now entered everyday life, with advanced tools and applications.
Technologies like:
● Hidden Markov Models (HMMs) – used in speech recognition and handwriting
recognition.
● Bayesian Networks – help AI make decisions under uncertainty.
● Machine Learning & Deep Learning – used in:
○ Chatbots
○ Self-driving cars
○ Recommendation systems (e.g., Netflix, YouTube)
Summary
Era Years Key Events
Gestation 1943–1955 McCulloch & Pitts' neuron model; Turing Test;
Dartmouth Conference
Early Enthusiasm 1952–1969 GPS, LISP, theorem provers, analogy solvers
Reality Check 1966–1973 Over-promising leads to AI Winter
Knowledge 1969–1979 DENDRAL and other expert systems
Systems
AI Industry Boom 1980–present Japan’s Fifth Gen Project; commercial expert
systems
Neural Networks 1986–present Rumelhart, Hinton, and deep learning models
Revival
Recent Advances 2000s–present HMMs, Bayesian models, widespread ML
adoption
Slide 7 - State of Art of AI
Current AI Innovations - Artificial Intelligence is no longer just theory
It is actively used in real-life applications across many industries.
Area Application / Example / Company
Autonomous Cars Tesla, [Link], Waymo, Apple, Hyundai-Kia, Audi,
Huawei
Speech Recognition Apple's Siri, Amazon Alexa, Google Assistant,
Microsoft Cortana
Chatbots Swelly, eBay, Lyft, Yes Sire, 1-800-Flowers
Search Engines Google Search using AI for better results (e.g.,
RankBrain)
Language Translation SYSTRAN, Google Translate
Natural Language Processing IBM Watson API, ChatGPT
(NLP)
Medical Diagnosis & Imaging AI in MRI scan interpretation, predictive analysis tools
Voice Assistants Embedded in mobile phones, smart speakers, smart
homes
Slide 8 - Practical Application of AI
● Autonomous Planning – NASA’s AI controlled a spacecraft without help.
● Game Playing – Deep Blue beat chess champ Kasparov.
● Self-Driving – ALVINN steered a van using computer vision.
● Medical Diagnosis – AI reads scans and suggests treatments.
● Logistics – DART planned military moves in hours, saving weeks.
● Robotic Surgery – AI guides precise operations (e.g., HipNav).
● Language & Puzzles – PROVERB solved crosswords using language rules.
Field Impact of AI
Healthcare Faster diagnosis, robotic surgery, patient
monitoring
Transport Self-driving cars, route optimization
Military Smart logistics, autonomous drones
Consumer Smart assistants, personalized recommendations
Electronics
Language Chatbots, translators, NLP-based tools
Education Personalized learning apps, virtual tutors
Finance Fraud detection, stock predictions, chatbots
Applications:
● Robotic Vehicles: Self-driving cars like STANLEY and BOSS navigate rough terrain and
urban streets autonomously, using sensors and AI to follow traffic rules and avoid
obstacles.
● Speech Recognition: AI systems handle customer service calls, recognizing speech
and managing conversations automatically.
● Autonomous Planning: Space missions (e.g., NASA's Remote Agent) use AI to
autonomously plan and manage spacecraft operations millions of miles from Earth.
● Game Playing: Programs like IBM’s Deep Blue have defeated world champions in
chess, showing strategic reasoning capabilities beyond most humans.
● Spam Filtering: AI filters billions of emails daily, identifying spam using learning
algorithms that adapt to new spamming tactics.
● Logistics Planning: AI tools like DART helped plan massive military deployments
efficiently, outperforming traditional manual methods.
● Robotics: Consumer robots like Roomba clean homes, while military robots like
PackBot handle dangerous tasks like bomb disposal.
● Machine Translation: AI translates languages using statistical models, even when
developers don’t speak the source language.
Intro:
[Link]
History:
[Link]
