The history of artificial intelligence is defined by cycles of extreme optimism followed by “winters” of reduced funding and interest. It has evolved from a theoretical branch of mathematics into a pervasive modern technology. 

The Foundations (Pre-1950)

Before AI was a formal field, it existed in science fiction and early mechanical concepts. 

• 1921: The term “robot” is coined by Karel Čapek in the play Rossum’s Universal Robots.
• 1943: Warren McCulloch and Walter Pitts publish the first mathematical model of a neural network.
• 1949: Edmund Berkeley’s book Giant Brains proposes that machines can think. 

The Birth of AI (1950–1956)

This era shifted AI from mythology to a serious academic discipline. 

• 1950: Alan Turing publishes “Computing Machinery and Intelligence,” introducing the Turing Test to measure machine intelligence.
• 1952: Arthur Samuel creates the first self-learning checkers program.
• 1955-1956: John McCarthy coins the term “Artificial Intelligence” during the proposal for the Dartmouth Workshop, which officially launched the field. 

The Golden Years & First AI Winter (1957–1979) 

Initial successes led to over-promising and a subsequent crash. 

• 1958: Frank Rosenblatt develops the Perceptron, the foundation for modern neural networks.
• 1966: Joseph Weizenbaum creates ELIZA, the first “chatterbot”.
• 1973-1974: The Lighthill Report in the UK and subsequent funding cuts by DARPA lead to the First AI Winter due to limited computing power and unmet expectations.

The Expert Systems Boom & Second Winter (1980–1993)

AI found commercial success through specialized knowledge bases before another decline. 

• 1980: XCON (expert configurer) becomes the first commercially successful expert system, saving Digital Equipment Corporation millions.
• 1981: Japan launches the Fifth Generation Computer project with $850 million to create human-level reasoning.
• 1987-1993: The Second AI Winter occurs as specialized AI hardware (Lisp machines) becomes obsolete compared to cheaper personal computers from Apple and IBM. 

The Age of Agents & Narrow AI (1993–2011) 

AI began achieving superhuman performance in specific, “narrow” tasks. 

• 1997: IBM’s Deep Blue defeats world chess champion Garry Kasparov.
• 2002: iRobot releases the Roomba, bringing autonomous AI into the home.
• 2011: IBM’s Watson wins Jeopardy! against human champions, and Apple releases Siri. 

The Deep Learning Revolution (2012–2021)

A massive surge in data and GPU power transformed the field. 

• 2012: AlexNet wins the ImageNet competition, proving the power of deep convolutional neural networks.
• 2016: Google DeepMind’s AlphaGo defeats world Go champion Lee Sedol.
• 2017: Researchers at Google propose the Transformer architecture, which becomes the backbone of modern large language models. 

The Generative AI Boom (2022–Present)

AI has entered the mainstream, moving toward Agentic AI that can plan and act autonomously. 

• 2022: OpenAI releases ChatGPT, sparking global interest in generative AI.
• 2023-2024: Focus shifts toward Multimodal AI (images, video, and text) and Agentic AI capable of completing complex workflows across multiple tools. 

AI history of artificial intelligence by era