What Is Artificial Intelligence?

Artificial intelligence is one of the most transformative technologies in human history. But what exactly is it? In this module, we'll trace AI's journey from a theoretical concept in the 1950s to the powerful systems shaping our world today.

Defining Artificial Intelligence

At its core, artificial intelligence is the science of creating machines that can perform tasks that typically require human intelligence. This includes understanding language, recognizing images, making decisions, and solving problems.

The field encompasses everything from simple rule-based systems ("if the temperature exceeds 100°F, turn on the cooling system") to sophisticated neural networks that can write code, compose music, and engage in nuanced conversation.

Key Distinction

AI is not a single technology — it's an umbrella term covering many approaches to making machines intelligent. Machine learning, deep learning, and large language models are all subfields within AI.

The Three Types of AI

AI researchers generally categorize AI into three levels based on capability:

Narrow AI (Weak AI)

We are here

AI designed to excel at a specific task. Every AI system you interact with today — ChatGPT, Siri, Tesla's autopilot, Netflix recommendations — is narrow AI. These systems can outperform humans at their specific task but cannot transfer that ability to other domains.

Artificial General Intelligence (AGI)

Emerging

A theoretical AI that could match human-level cognitive ability across any intellectual task — learning, reasoning, adapting, and transferring knowledge between domains just like a human. Some researchers believe we may be approaching early forms of AGI, but this remains hotly debated.

Artificial Superintelligence (ASI)

Theoretical

AI that surpasses human intelligence in virtually every domain — scientific creativity, social intelligence, and general wisdom. ASI remains a theoretical concept and is the focus of much of the existential risk research in AI safety.

The AI Timeline: Key Milestones

Understanding where AI came from helps us understand where it's going. Here are the defining moments that shaped the field:

1950

The Turing Test

Alan Turing publishes 'Computing Machinery and Intelligence,' proposing a test for machine intelligence that remains influential today.

1956

AI Is Born

The Dartmouth Conference coins the term 'artificial intelligence.' Researchers optimistically predict human-level AI within a generation.

1974–1980

First AI Winter

The 1973 Lighthill Report and mounting disappointment trigger funding cuts. Initial hype fades as AI systems fail to live up to promises.

1997

Deep Blue Defeats Kasparov

IBM's Deep Blue defeats world chess champion Garry Kasparov, demonstrating AI's potential in strategic reasoning.

2012

Deep Learning Revolution

AlexNet wins ImageNet by a massive margin, proving deep neural networks can dramatically outperform traditional methods. The modern AI era begins.

2016

AlphaGo Moment

Google DeepMind's AlphaGo defeats world Go champion Lee Sedol — a game previously thought too complex for AI. Viewed by 200 million people.

2017

Attention Is All You Need

Google researchers publish the Transformer paper, introducing the architecture that would power GPT, BERT, and every modern LLM.

2022

ChatGPT Changes Everything

OpenAI launches ChatGPT, reaching 100 million users in two months. AI enters mainstream consciousness overnight.

2023–2024

The Foundation Model Era

GPT-4, Claude, Gemini, Llama, and Mistral push boundaries. AI coding, reasoning, and multimodal capabilities advance rapidly.

2025–2026

The Age of Agents

AI systems gain the ability to use tools, browse the web, write and execute code, and complete complex multi-step tasks autonomously.

From Rule-Based to Foundation Models

AI has evolved through several distinct paradigms, each building on the last:

Era	Approach	How It Works	Example
1950s–1980s	Rule-Based / Expert Systems	Humans write explicit rules	Medical diagnosis systems
1990s–2000s	Classical Machine Learning	Algorithms learn from structured data	Spam filters, recommendation engines
2012–2020	Deep Learning	Neural networks learn from massive datasets	Image recognition, speech-to-text
2020–Present	Foundation Models / LLMs	Massive models trained on internet-scale data, adaptable to many tasks	ChatGPT, Claude, Gemini

Think of it this way

Rule-based systems are like following a recipe exactly. Machine learning is like learning to cook by tasting thousands of dishes. Foundation models are like having a chef who has read every cookbook ever written and can improvise any cuisine on demand.

Recommended Resources

Video

But what is a neural network?

3Blue1Brown

The gold-standard visual explainer of how neural networks work. Beautiful animations make abstract concepts intuitive.

Video

How AI Could Empower Any Business

TED (Andrew Ng)

Andrew Ng's accessible TED talk on how AI can empower any business. Perfect introduction to AI for non-technical audiences.

Course

Elements of AI

University of Helsinki

Free, comprehensive online course covering AI fundamentals. Over 1 million students enrolled worldwide.

Key Takeaways

1AI is an umbrella term — it encompasses rule-based systems, machine learning, deep learning, and foundation models.
2All commercially deployed AI today is 'narrow AI' — excellent at specific tasks but not generally intelligent.
3The Transformer architecture (2017) is the foundation of modern LLMs like ChatGPT, Claude, and Gemini.
4The field has accelerated dramatically since 2012, with major breakthroughs happening on a yearly or even monthly basis.
5Understanding AI's history helps you evaluate hype vs. reality when making decisions about AI adoption.