Differentiate between narrow and general AI

Aspect Narrow AI (Weak AI) General AI (Strong AI)

AI capable of performing any

Definition AI designed for a specific task.
intellectual task like a human.

Scope Narrow and limited to one domain. Broad, across multiple domains.

Learning Can learn, adapt, and transfer

Learns only for predefined tasks.
Ability knowledge to new tasks.

Voice assistants, spam filters,

Still theoretical – true human-like
Examples recommendation systems, face
AI.
recognition.

Level of Human-level or higher

Task-specific intelligence.
Intelligence intelligence.

Can solve unseen problems and

Flexibility Cannot perform beyond its training.
think creatively.

Not yet developed, under

Existence Exists today, widely used.
research.

Self- Expected to have reasoning, self-

No self-awareness or consciousness.
awareness awareness, and understanding.
Applications of AI in Healthcare

1. Medical Imaging and Diagnostics

AI analyzes X-rays, CT scans, and MRI images to detect diseases like tumors, fractures, and
pneumonia.

Helps in early detection of cancer, brain disorders, and eye diseases.

2. Drug Discovery and Development

AI models predict how drugs will interact with the human body.

Reduces the time and cost of developing new medicines.

3. Virtual Health Assistants

Chatbots and AI assistants provide medical information, remind patients about medicines,
and answer basic health queries.

4. Predictive Analytics

AI analyzes patient data to predict disease risks (like heart attack or diabetes).

Helps in preventive healthcare.

5. Robotic Surgery

AI-powered robots assist surgeons with precision and minimal invasion.

Reduces errors, pain, and recovery time.

6. Personalized Treatment

AI suggests customized treatment plans based on a patient’s genetics, lifestyle, and

medical history.
 7. Hospital Management

AI is used in scheduling, patient record management, and resource optimization.

Improves efficiency and reduces waiting times.

8. Remote Patient Monitoring (Telemedicine)

AI analyzes data from wearable devices (like smartwatches, glucose monitors, heart-rate
trackers).

Doctors can monitor patients in real-time.

Components of AI Programs

1. Knowledge Base

Stores facts, rules, and relationships about the world.

Example: Medical database, expert system rules.

2. Inference Engine

Logical reasoning mechanism that applies rules to known facts to derive new information.
Example: Diagnosis system deciding treatment based on symptoms.

3. Learning Component

Enables AI to improve performance from past experiences or data (Machine Learning, Deep
Learning).

Example: AI model learning to recognize images better over time.

4. Natural Language Processing (NLP)

Helps AI understand, process, and generate human language.

Example: Chatbots, voice assistants.

5. Perception Component

Uses sensors or input devices to perceive the environment (vision, speech, touch).
Example: Computer vision in self-driving cars.

6. Problem-Solving Component

Provides strategies to find solutions (search algorithms, optimization methods).

Example: Route planning in GPS.

7. Planning and Decision-Making

AI decides future actions based on goals, constraints, and predictions.

Example: Robot planning steps to reach a destination.

8. User Interface

Allows interaction between humans and AI systems.

Example: Dashboard, voice commands, chatbot interface.

Evolution and History of AI

1. Early Foundations (1940s – 1950s)

1943: McCulloch & Pitts developed the first mathematical model of a neuron.

1950: Alan Turing introduced the Turing Test to check if a machine can “think”.

1951: The first AI programs for checkers and chess were written.

2. Birth of AI (1956)

1956: The term Artificial Intelligence was coined at the Dartmouth Conference by John
McCarthy.

Early programs like Logic Theorist and General Problem Solver (GPS) were developed.

3. Golden Years (1956 – 1974)

AI research expanded into problem-solving, reasoning, and games.

Development of LISP programming language (by John McCarthy) for AI.

Optimism led to high funding and expectations.

4. First AI Winter (1974 – 1980)

Funding reduced due to limited computing power and over-promised results.

AI research slowed down.

5. Expert Systems Era (1980 – 1987)

AI revived with Expert Systems (rule-based programs).

Example: MYCIN (medical diagnosis), DENDRAL (chemical analysis).

Commercial use of AI in industries.

 6. Second AI Winter (1987 – 1993)

Expert systems became costly and inflexible.

Market interest and funding declined again.

7. Modern AI & Machine Learning (1993 – 2010)

Rise of Machine Learning and Data-driven AI.

Use of algorithms like Neural Networks, Support Vector Machines.

AI applied in speech recognition, handwriting, and recommendation systems.

8. Deep Learning & Big Data Era (2010 – Present)

Advances in Deep Learning, GPUs, and Big Data boosted AI capabilities.

AI achieved human-level performance in vision, speech, and games.

Breakthroughs:

2011: IBM Watson won Jeopardy!

2012: Deep learning revolution (ImageNet competition).

2016: AlphaGo defeated world champion in Go.

Current AI: self-driving cars, medical diagnosis, robotics, ChatGPT, etc.

In summary:

1940s–1950s → Foundations.

1956 → Birth of AI (Dartmouth).

1960s–70s → Growth and optimism.

1970s & 1980s → Two AI Winters.

1980s → Expert systems.

1990s–2000s → Machine learning.

2010s–Present → Deep learning & modern AI applications.

Categorization of Intelligent Systems

Artificial Intelligence (AI) systems can be categorized based on their capabilities and
functionalities.

A. Based on Capabilities

1. Narrow AI (Weak AI)

Designed for a specific task only.

Cannot perform beyond its programmed scope.

Examples:
Google Translate (language translation)

Siri, Alexa (voice assistants)

Spam email filters

2. General AI (Strong AI)

Performs any intellectual task like a human.

Can learn, adapt, and apply knowledge in multiple domains.

Examples:

Still theoretical, not yet developed.

A robot with human-like intelligence (future goal).

3. Super AI
Hypothetical AI that surpasses human intelligence in all aspects.

Would have self-awareness, creativity, and decision-making beyond humans.

Examples:

Does not exist yet (conceptual).

B. Based on Functionalities

1. Reactive Machines

Work only on present input.

No memory, cannot use past experiences.

Examples: IBM’s Deep Blue (chess-playing computer).

 2. Limited Memory Systems

Can use past data for a short period.

Widely used in current AI applications.

Examples:

Self-driving cars (use past traffic data for decision-making).

Virtual assistants (store recent interactions).

3. Theory of Mind AI

Future AI that can understand emotions, intentions, and social interactions.

Examples: Still in research stage.

4. Self-Aware AI

Advanced future AI that has consciousness and self-awareness.

Examples: Only theoretical, not developed yet.

8 Real-World Use Case of AI (with Diagram)

Use Case: AI in Healthcare – Medical Imaging Diagnosis

Problem: Detecting diseases (like tumors, fractures, pneumonia) from X-rays, MRI, CT
scans.

AI Solution: Deep Learning algorithms analyze medical images, compare patterns, and
assist doctors in early detection.

Benefits: Faster diagnosis, higher accuracy, reduces human error.

Diagram (Conceptual Flow):

9 Architecture of an Intelligent Agent

An Intelligent Agent is a system that perceives the environment through sensors and acts
upon it using actuators.

Components of Intelligent Agent Architecture

1. Sensors → Collect input from the environment (e.g., camera, microphone).

2. Perception/Knowledge Base → Stores information and past data.

3. Inference/Reasoning Engine → Analyzes data, applies rules, and makes decisions.

4. Learning Component → Improves performance using experience (Machine

Learning).

5. Actuators → Perform actions in the environment (e.g., robot arm, display, speaker).
 10 NLP, Robotics, and ML as Subfields of AI

1. Natural Language Processing (NLP)

Deals with the interaction between computers and human language.

Applications: Chatbots, Google Translate, Voice Assistants (Alexa, Siri).

Tasks: Speech recognition, sentiment analysis, machine translation.

2. Robotics

Branch of AI that designs and builds intelligent machines capable of performing tasks in
the physical world.

Applications: Industrial robots, self-driving cars, surgical robots, drones.

Involves perception (sensors), motion planning, and control.

3. Machine Learning (ML)

Subfield of AI that enables systems to learn from data and improve without explicit
programming.

Types: Supervised, Unsupervised, Reinforcement Learning.

Applications: Spam detection, recommendation systems, fraud detection, image

recognition.

Sub-areas of Artificial Intelligence (AI)

AI is a broad field that consists of several specialized sub-areas:

1. Machine Learning (ML)

Enables computers to learn from data and improve performance without explicit
programming.

Types: Supervised, Unsupervised, Reinforcement learning.

Applications: Spam detection, fraud detection, recommendation systems.

 2. Natural Language Processing (NLP)

Focuses on interaction between computers and human language.

Tasks: Speech recognition, machine translation, sentiment analysis.

Applications: Google Translate, chatbots, virtual assistants.

3. Robotics

Combines AI with mechanical engineering to create intelligent machines.

Robots use sensors (vision, touch, sound) and actuators to perform tasks.

Applications: Industrial robots, self-driving cars, medical robots, drones.

4. Computer Vision

Enables machines to “see” and interpret visual data from images or videos.
Tasks: Object detection, facial recognition, motion tracking.

Applications: CCTV surveillance, medical imaging, autonomous vehicles.

5. Expert Systems

AI programs that use a knowledge base + inference engine to provide solutions.

Applications: Medical diagnosis (MYCIN), chemical analysis (DENDRAL).

6. Speech Recognition & Processing

Converts spoken language into text and interprets meaning.

Applications: Siri, Alexa, Google Assistant, call center automation.

 7. Planning & Decision Making

AI systems that set goals, plan actions, and optimize resources to achieve outcomes.

Applications: GPS route planning, supply chain optimization, robotics path planning.

8. Neural Networks & Deep Learning

Models inspired by the human brain, used for pattern recognition and complex decision-
making.

Applications: Image classification, fraud detection, self-driving cars.