History of Artificial Intelligence

Turing publishes 'Computing Machinery and Intelligence' in Mind journal. Introduces the Imitation Game (later known as the Turing Test) — a benchmark for machine intelligence that shaped AI research for decades.

Dartmouth Conference: John McCarthy, Marvin Minsky, Claude Shannon, and Nathaniel Rochester coin the term 'Artificial Intelligence.' The workshop defines AI as a formal academic discipline. Funding and optimism surge.

Frank Rosenblatt at Cornell builds the Mark I Perceptron, the first machine capable of learning through trial and error. The Navy funds the project. The New York Times reports a computer that 'will be able to walk, talk, see, write, reproduce itself and be conscious of its existence.'

SRI International builds Shakey, the first general-purpose mobile robot that reasons about its own actions. It combines computer vision, route planning, and problem solving. Shakey inspires decades of robotics and autonomous systems research.

Rule-based expert systems like MYCIN (medical diagnosis) and XCON (computer configuration) prove commercially viable. Corporations invest billions. Japan launches the Fifth Generation Computer project.

Rumelhart, Hinton, and Williams publish 'Learning representations by back-propagating errors' in Nature. The paper provides an efficient method to train multi-layer neural networks, resurrecting a field that Minsky and Papert had nearly killed in 1969.

Tim Berners-Lee publishes the first website, launching the World Wide Web. Within a decade, billions of web pages create the massive text and image datasets that later fuel machine learning. No web, no big data; no big data, no modern AI.

IBM's Deep Blue beats Garry Kasparov in a six-game match. First time a computer defeats a reigning world champion under standard rules. The event captivates the world but uses brute force, not learning.

Geoffrey Hinton publishes breakthrough work on deep belief networks, reviving neural network research. The term 'deep learning' enters mainstream AI vocabulary. GPUs make training large networks feasible.

Netflix offers $1 million to anyone who can improve its recommendation algorithm by 10%. The competition attracts 40,000 teams from 186 countries over three years. BellKor's Pragmatic Chaos wins in 2009, proving collaborative filtering at scale.

Apple launches Siri with the iPhone 4S, making voice-activated AI a mass-market consumer product. Originally a DARPA-funded SRI International project, Siri normalizes the idea of talking to machines for hundreds of millions of users.

Ian Goodfellow introduces Generative Adversarial Networks in a landmark paper. Two neural networks compete — one generates, one discriminates — producing increasingly realistic synthetic data. GANs become the foundation of AI image generation.

DeepMind's AlphaGo beats world champion Lee Sedol 4-1 in a game long considered too complex for computers. Move 37 in Game 2 — a play no human would make — becomes iconic. 200 million people watch live.

Google researchers publish the Transformer architecture paper. Self-attention mechanism replaces recurrence. This single paper becomes the foundation of GPT, BERT, and every major language model that follows.

Waymo One becomes the first commercial autonomous ride-hailing service, launching in Phoenix, Arizona. Passengers ride without a safety driver. By 2024, the service expands to San Francisco and Los Angeles, logging millions of autonomous miles.

OpenAI announces GPT-2 with 1.5B parameters. Initially withholds the full model citing misuse concerns. The decision sparks global debate about AI safety, open-source ethics, and who controls powerful AI.

DeepMind's AlphaFold2 predicts protein structures with atomic accuracy, solving a 50-year-old grand challenge of biology. Later releases structures for 200M+ proteins, accelerating drug discovery worldwide.

GitHub and OpenAI launch Copilot, an AI pair programmer trained on billions of lines of public code. Within two years, it generates 46% of code in files where it is enabled. Software development begins its biggest workflow shift since Stack Overflow.

Stability AI releases Stable Diffusion as open-source, allowing anyone to generate images locally. Millions download it within weeks. The decision democratizes generative AI and ignites debate over copyright, artist rights, and model governance.

OpenAI launches ChatGPT. Reaches 100 million users in 2 months — the fastest-growing consumer application in history. AI becomes a daily tool for millions. The conversation shifts from 'if' to 'how fast.'

OpenAI releases GPT-4 — processes text and images, passes the bar exam in the 90th percentile, and demonstrates reasoning capabilities that surprise even its creators. The gap between AI and human cognition narrows visibly.

European Parliament passes the AI Act — the world's first comprehensive legal framework for artificial intelligence. Risk-based approach: bans social scoring, regulates high-risk systems, requires transparency for generative AI.

Anthropic's Claude, Google's Gemini, Meta's Llama, and Mistral push capabilities forward. Open-source models close the gap with proprietary ones. AI becomes commoditized — the question shifts to trust, safety, and alignment.

Geoffrey Hinton shares the Nobel Prize in Physics for foundational work on artificial neural networks. Days later, Demis Hassabis receives the Nobel Prize in Chemistry for AlphaFold's protein structure predictions. AI research earns science's highest honors for the first time.