Detailed · 30 events
A History of Artificial Intelligence
1950s
02 events
null0 (Wikimedia Commons, via Flickr) · CC BY-SA 2.0 · Commons ↗ A funding proposal dated 31 August 1955 and signed by John McCarthy, Marvin Minsky, Nathaniel Rochester, and Claude Shannon put the phrase 'artificial intelligence' into an official research document; the workshop it asked for ran at Dartmouth College in New Hampshire the following summer. The '2 month, 10 man' figure was the plan, not a roster — attendance was fluid and daily sessions drew three to eight people. Its historical weight lies less in any specific technical result than in establishing the field's own name and the researcher network that would carry it for decades.
- Related people
- John McCarthy · Marvin Minsky · Claude Shannon
- Related organizations
- International Business Machines (IBM)
- Appears in
- A General History of Information Technology
John C. Hay, Albert E. Murray — Mark I Perceptron Operators' Manual, Cornell Aeronautical Laboratory (Wikimedia Commons) · Public domain · Commons ↗ In July 1958 the US Office of Naval Research held a press conference for the perceptron of Frank Rosenblatt of the Cornell Aeronautical Laboratory: a learning device that weights its inputs, applies a threshold, and adjusts the weights whenever it errs. The demonstration was a simulation running on an IBM 704 that learned, in about fifty trials, to tell cards marked on the left from cards marked on the right. The theory appeared the same year in Psychological Review 65(6), 386–408. The Mark I Perceptron—a separate, purpose-built electromechanical machine whose sensory layer was an array of 400 photocells in a 20×20 grid—was assembled at the same laboratory. It marks the start of the neural-network lineage and remains the structural skeleton beneath the MLP, the convolutional network, and modern deep learning.
- Related people
- Frank Rosenblatt
- Related terms
- Deep Learning · Neural Network · Supervised Learning
1960s
01 events
Joseph Weizenbaum (via elizagen.org), Wikimedia Commons · CC0 · Commons ↗ Joseph Weizenbaum wrote ELIZA in MAD-SLIP at MIT and published it in Communications of the ACM in January 1966. ELIZA was the engine; what it said was governed by a swappable 'script', and the famous one—DOCTOR—imitated a Rogerian psychotherapist. It picked keywords out of the input and rephrased them back using decomposition and reassembly rules. In the paper Weizenbaum argued the impression of being understood is the speaker's own contribution—people attribute 'background knowledge, insights and reasoning ability' to their conversational partner—and he later wrote *Computer Power and Human Reason* (1976) to warn against the effect.
- Related people
- Joseph Weizenbaum
1970s
01 events
University of Edinburgh, School of Informatics, Freddy II Project · CC BY 4.0 · Commons ↗ Commissioned by the British Science Research Council, James Lighthill produced a report dated July 1972 and published in 1973 that concluded 'in no part of the field have the discoveries made so far produced the major impact that was then promised.' The SRC withdrew most of its AI funding, leaving programmes standing only at Edinburgh, Essex and Sussex. The American retrenchment was a separate chain of events: the National Research Council's ALPAC report (1966) ended machine-translation funding, the Mansfield Amendment (1969) obliged DARPA to back mission-oriented work, and in 1974 DARPA cancelled its Speech Understanding Research programme. The decade that followed — lean on both sides of the Atlantic, for its own reasons on each — is known as the first AI winter, the first sustained rebuttal of the optimistic schedules set out by Newell, Minsky, and others.
1980s
01 events
Jason Riedy (Wikimedia Commons) · CC BY 2.0 · Commons ↗ Stanford's MYCIN (bacterial-infection diagnosis) and DEC's XCON (automated VAX configuration) showed the appeal of encoding domain expertise as explicit 'if-then' rules. The approach spread rapidly into industry in the early 1980s; Japan's MITI launched the Fifth Generation Computer Project (1982–92). By 1985, corporate spending on AI had passed US$1 billion a year. The maintenance burden and brittleness of rule systems, however, collapsed the boom—the LISP machine market fell apart in 1987—and the second AI winter followed.
1990s
01 events
Christina Xu (Wikimedia Commons, via Flickr) · CC BY 2.0 · Commons ↗ In a six-game rematch played in New York from 3 to 11 May 1997, IBM's chess-specific machine Deep Blue beat the reigning world champion Garry Kasparov 3½–2½. Their first match, in Philadelphia in February 1996, had gone to Kasparov 4–2—though Deep Blue took game 1 there, the first game a computer had won against a reigning world champion under regular time controls. Over the intervening year IBM strengthened the endgame databases and the evaluation function and brought in grandmaster advisers. The 1997 result was the first time a computer won a match against a reigning world champion under standard tournament time controls. Its substance was game-tree search plus a specialised evaluation function, with no modern learning involved.
- Related organizations
- International Business Machines (IBM)
- Appears in
- A General History of Information Technology
2010s
05 events- EVT.007T2iPhone 4S and Siri — Jobs' Last LaunchAppears inA History of the iPhoneA History of Mobile Phones and Smartphones
Daniel Voigt Godoy (Wikimedia Commons) · CC BY 4.0 · Commons ↗ Entering as 'SuperVision', Alex Krizhevsky, Ilya Sutskever, and Geoffrey Hinton of the University of Toronto took the ILSVRC 2012 classification task with a top-5 error rate of 15.3% (16.4% on supplied training data alone). The best result from any other team was 26.2%, from the University of Tokyo's ISI group using conventional methods. Their convolutional network—about 60 million parameters, trained for five to six days on two NVIDIA GTX 580 3GB GPUs—became known as 'AlexNet' after its first author. Computer vision moved from hand-engineered features to deep learning from that point forward.
- Related people
- Geoffrey Hinton · Ilya Sutskever
- Related organizations
- NVIDIA
- Related terms
- Deep Learning · Graphics Processing Unit (GPU) · Neural Network
- Appears in
- A General History of Information Technology · A History of Semiconductors and Hardware
Immigrant laborer (Wikimedia Commons) · CC0 1.0 (public domain dedication) · Commons ↗ DeepMind's AlphaGo won its five-game series against Lee Sedol, one of the world's top Go players, 4–1, played in Seoul between 9 and 15 March 2016. The combinatorial explosion of Go had long been thought to put a machine victory at least a decade away; the combination of Monte Carlo tree search and deep reinforcement learning broke that wall. In Game 2, AlphaGo's move 37 — a fifth-line shoulder hit the program itself rated a one-in-ten-thousand human choice — was so unlike anything a professional would have played that the Go community discussed it for weeks. In Game 4, the single game Lee won, he answered with move 78, a wedge of equally long odds that commentators named the 'divine move'.
- Related people
- Demis Hassabis
- Related organizations
- DeepMind · Google
- Related terms
- Reinforcement Learning
- Appears in
- A General History of Information Technology
Yuening Jia (Wikimedia Commons), DOI:10.1088/1742-6596/1314/1/012186 · CC BY-SA 3.0 · Commons ↗ Ashish Vaswani and seven colleagues at Google Brain and Google Research proposed the Transformer—a sequence-to-sequence architecture built solely on self-attention. Posted to arXiv on 12 June 2017, it was presented that December at NIPS 2017. It displaced the RNN and LSTM, the prevailing NLP architectures, with a structure that parallelised easily and trained efficiently. By 2026 the paper had been cited more than 250,000 times, placing it among the ten most-cited papers of the twenty-first century; every modern large language model (BERT, the GPT family, Claude, Gemini) is a descendant.
- Related organizations
- Related terms
- Large Language Model (LLM) · Transformer
Daniel Voigt Godoy (Wikimedia Commons) · CC BY 4.0 · Commons ↗ Jacob Devlin and colleagues at Google AI posted BERT (Bidirectional Encoder Representations from Transformers) to arXiv as 1810.04805. Pretraining a Transformer encoder bidirectionally as a masked-language model set new state of the art on eleven NLP tasks; BERT-Large scored 80.5 on the official GLUE leaderboard against GPT's 72.8. The paper was presented at NAACL 2019, where it won Best Long Paper. It established the pretrain-then-fine-tune paradigm that underlies every modern LLM, and with the autoregressive GPT family forms one of the two great currents of Transformer-based language modelling.
- Related organizations
- Related terms
- Fine-tuning · Transformer · Token · Embedding
- Appears in
- A History of Search Engines
2020s
19 events
OpenAI (Wikimedia Commons) · Public domain (below threshold of originality) · Commons ↗ OpenAI's 175-billion-parameter language model. The paper, "Language Models are Few-Shot Learners," went up on arXiv on 28 May 2020; on 11 June an API running models from the same family opened as a waitlisted private beta. At more than a hundred times the scale of GPT-2 (1.5B), it showed 'few-shot' competence—translation, summarisation, question answering, and code produced from a handful of in-prompt examples with no gradient updates. The paper was one of three NeurIPS 2020 Best Paper Award winners, and it pushed the research conversation toward scaling large language models.
- Related organizations
- OpenAI
- Related terms
- Large Language Model (LLM)
OpenAI (Wikimedia Commons) · Public domain (below threshold of originality) · Commons ↗ OpenAI released ChatGPT as a free research preview—a model from the GPT-3.5 series, which finished training in early 2022, fine-tuned for dialogue with RLHF and reachable from any browser. One million users in five days (Sam Altman's figure); an estimated one hundred million monthly active users within two months—the latter a UBS estimate built on Similar Web data and reported by Reuters on 1 February 2023, not a number OpenAI published. It was the fastest ramp any consumer application had then managed (Meta's Threads passed 100 million sign-ups in five days in July 2023), and it moved generative AI overnight from specialist research into homes, schools, and workplaces, forcing strategic pivots at Google, Meta, Anthropic, and Microsoft.
- Related people
- Sam Altman
- Related organizations
- OpenAI · Microsoft Corporation
- Related terms
- Fine-tuning
- Appears in
- A General History of Information Technology
Hstoops / Microsoft (Wikimedia Commons) · Public domain (below threshold of originality; trademark applies) · Commons ↗ Microsoft unveiled 'New Bing'—Bing search with a chat experience built in—as a waitlisted limited preview. On the day, the model was described only as 'a new, next-generation OpenAI large language model that is more powerful than ChatGPT'; Microsoft confirmed it was GPT-4 on 14 March. Microsoft 365 Copilot followed on 16 March, bringing LLMs into the central battlefronts of search and productivity software at once. Coming barely two months after the ChatGPT launch, it triggered the industry-wide pivot to 'how do we integrate an LLM into our existing product?'. Google had announced Bard the day before, on 6 February, and the factual error in its promo clip surfaced on the 8th.
- Related organizations
- Microsoft Corporation · OpenAI
- Related terms
- Large Language Model (LLM)
- Appears in
- A History of Microsoft · A History of Search Engines
Anthropic (Wikimedia Commons) · Public domain (below threshold of originality) · Commons ↗ Anthropic—founded in January 2021 by seven people who left OpenAI, among them the Amodei siblings Dario and Daniela—began offering Claude and the lighter Claude Instant. This was not an open release but request-based access following a closed alpha with Notion, Quora, and DuckDuckGo. OpenAI announced GPT-4 the same day. Anthropic's differentiator was Constitutional AI, set out in a December 2022 paper: human oversight enters only through a written list of principles, with the model's own critiques, revisions, and RL from AI Feedback doing the rest. Google invested about US$300 million in February 2023; Amazon announced up to US$4 billion in September 2023 and later doubled it to US$8 billion; in April 2026 Google pledged up to a further US$40 billion.
- Related organizations
- Anthropic · OpenAI · Google
- Related terms
- Large Language Model (LLM) · Reinforcement Learning
- Appears in
- A General History of Information Technology
Bubeck, S. et al., "Sparks of Artificial General Intelligence" (arXiv:2303.12712), Microsoft Research — via Wikimedia Commons · CC BY 4.0 · Commons ↗ OpenAI announced GPT-4, a multimodal model taking image and text inputs and producing text output—though image input was a research preview at announcement and did not reach general availability until September 2023. The technical report gives a simulated Uniform Bar Exam score of 298/400, around the top 10% of test takers. That 90th-percentile figure was later shown to be inflated: measured against a July sitting of the same exam GPT-4 falls below the 69th percentile, drops to roughly the 48th percentile against those who passed, and to roughly the 15th percentile on the essays (Martínez, Artificial Intelligence and Law, 2024). Its AP results were uneven—5s in Biology, Environmental Science and Macroeconomics, but 2s in both AP English Language (14th–44th percentile) and AP English Literature (8th–22nd), below a passing grade. Parameter count, hardware, training compute, and dataset construction were all withheld, citing "the competitive landscape and the safety implications of large-scale models"—the generation that marked OpenAI's turn from the publish-and-research posture of GPT-3 to a commercially closed lab. GPT-4 was retired from ChatGPT on 30 April 2025 in favour of GPT-4o, remaining available in the API.
- Related organizations
- OpenAI
- Related terms
- Large Language Model (LLM)
Christopher P. Michel (Wikimedia Commons) · CC BY-SA 4.0 · Commons ↗ Geoffrey Hinton—called the 'Godfather of AI'—revealed in a New York Times interview that he had left Google after a decade so he could speak about the dangers of AI without considering how it affected Google. He said he regretted parts of his life's work, and that what he had thought was 30 to 50 years off he no longer believed was. He warned of a flood of misinformation, job displacement, and autonomous weapons. An exceptional alarm from the man whose 2012 AlexNet ignited the deep-learning revolution, it accelerated international debate on AI regulation.
- Related people
- Geoffrey Hinton
- Related organizations
- Related terms
- Deep Learning
- Appears in
- A General History of Information Technology
Adam Schultz / The White House (Wikimedia Commons) · Public domain (PD-USGov-POTUS) · Commons ↗ US President Joe Biden signed an executive order titled 'Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence' (Executive Order 14110). It required developers of models above a compute threshold to report to the government, directed NIST to lead the development of AI safety evaluation standards, and addressed discrimination in immigration, housing, and criminal justice. The most comprehensive US AI regulatory framework to date. The Trump administration would rescind it in January 2025.
Steve Jennings / TechCrunch (Wikimedia Commons) · CC BY 2.0 · Commons ↗ On 17 November the OpenAI board removed Sam Altman as CEO, its announcement saying he 'was not consistently candid in his communications with the board'. CTO Mira Murati was named interim CEO, then replaced on 20 November by former Twitch CEO Emmett Shear. That same day more than 700 of roughly 770 employees — 745 by the final count — signed a letter demanding the board resign, and Microsoft said it would take Altman and Greg Brockman aboard. An agreement in principle to reinstate Altman was announced late on 21 November US Pacific time and finalised on 29 November, with an initial board of Bret Taylor (chair), Larry Summers, and the sole holdover Adam D'Angelo. The episode is remembered as a shock to the very governance structure of the AI industry.
- Related people
- Sam Altman · Ilya Sutskever
- Related organizations
- OpenAI · Microsoft Corporation
- Appears in
- A General History of Information Technology
Electronic Frontier Foundation (Wikimedia Commons) · CC BY 3.0 · Commons ↗ On 10 January, OpenAI quietly removed the 'military and warfare' prohibition from its usage policy, opening the door to defense work; The Intercept reported it on 12 January. On 26 June Anthropic wrote contractual exceptions to its usage policy for selected government agencies, and on 7 November it announced with Palantir and AWS that Claude 3 and 3.5 would run in Palantir's IL6-accredited classified environment. The pretense that generative AI was a peaceful Silicon Valley tool fell away; AI was openly recast as critical national-security infrastructure. The coupling ran on through the Pentagon's 2025 CDAO awards of up to $200 million each to four labs — and then broke open in February 2026, when Anthropic's refusal to drop its bans on fully autonomous weapons and mass domestic surveillance put it in direct conflict with the administration.
- Related organizations
- OpenAI · Anthropic · Google · Amazon
- Appears in
- A General History of Information Technology
European Parliament / Laia Ros (Wikimedia Commons) · CC BY 2.0 · Commons ↗ The European Parliament adopted the AI Act by 523 to 46 with 49 abstentions. It classifies AI systems into four risk tiers—unacceptable, high, limited, minimal—prohibits some outright (social scoring, untargeted facial-image scraping, real-time remote biometric identification for law enforcement), and imposes strict duties on high-risk AI in medicine, hiring, and justice. General-purpose AI (GPAI) models owe transparency and copyright compliance, with extra duties above a 10²⁵ FLOP training-compute presumption. The Council adopted it on 21 May, it appeared in the Official Journal on 12 July as Regulation (EU) 2024/1689, and it entered into force on 1 August 2024. The Commission calls it the world's first comprehensive legal framework on AI—but the Digital Omnibus of July 2026, Regulation (EU) 2026/1744, deferred the high-risk obligations to 2 December 2027 and 2 August 2028.
- EVT.022T2iPhone 16 and Apple IntelligenceAppears inA History of the iPhoneA History of Mobile Phones and Smartphones
MisterLahey (Wikimedia Commons) · CC BY-SA 4.0 · Commons ↗ Constellation Energy announced a 20-year power purchase agreement with Microsoft that will restart Three Mile Island Unit 1, shut down for economic reasons on 20 September 2019. It is the largest PPA in Constellation's history and returns roughly 835 megawatts to the grid. The point most often got wrong belongs first: the reactor that failed in 1979 was Unit 2, a separate facility. Constellation's own announcement states that "TMI Unit 1 is a fully independent facility, and its long-term operation was not impacted by the Unit 2 accident." The plant was renamed the Crane Clean Energy Center after the late Chris Crane. The restart was announced for 2028 and later pulled forward to 2027 after PJM expedited the interconnection approval; in November 2025 the Department of Energy closed a US$1 billion loan. The backdrop is the documented growth in data centre electricity demand published by the IEA and the US Department of Energy.
- Related organizations
- Microsoft Corporation
- Appears in
- A General History of Information Technology · A History of Cloud Computing
Jay Dixit (Wikimedia Commons) · CC BY-SA 4.0 · Commons ↗ On 8 October, the Nobel Prize in Physics went in equal shares to John Hopfield (the Hopfield network) and Geoffrey Hinton (the Boltzmann machine), 'for foundational discoveries and inventions that enable machine learning with artificial neural networks'. The following day brought a separate prize: the Chemistry Nobel, one half to David Baker 'for computational protein design' and the other half jointly to Google DeepMind's Demis Hassabis and John Jumper 'for protein structure prediction' (AlphaFold2). Both basic-science prizes going to AI work in the same week was unprecedented—a recognition of machine learning as an established scientific field.
- Related people
- Geoffrey Hinton · Demis Hassabis
- Related organizations
- DeepMind
- Related terms
- Deep Learning · Neural Network
- Appears in
- A General History of Information Technology
DeepSeek (Wikimedia Commons) · MIT License · Commons ↗ China's DeepSeek released DeepSeek-R1 and DeepSeek-R1-Zero, open-weights reasoning models comparable to OpenAI's o1-1217, under the MIT licence. What shocked the market was the cost line in the technical report for their base model, DeepSeek-V3 (December 2024): US$5.576 million for the final training job, on the paper's own assumption of $2 per H800 GPU-hour — and the H800 was the down-specified part built to satisfy US export controls on China. On 27 January NVIDIA stock fell about 17%, erasing roughly US$589 billion in market value, the largest single-day loss for any US company. The narrative of a US monopoly on frontier LLMs cracked under the combination of low cost, open weights, and a Chinese provenance.
- Related organizations
- NVIDIA
- Related terms
- Graphics Processing Unit (GPU) · Reinforcement Learning · Fine-tuning
- Appears in
- A General History of Information Technology
Daniel Torok / The White House (Wikimedia Commons) · Public domain (PD-USGov-POTUS) · Commons ↗ On 20 January the second Trump administration took office and Executive Order 14148 revoked 67 Biden orders, EO 14110 among them. The next day the White House hosted the announcement of the Stargate Project: a new company backed by SoftBank, OpenAI, Oracle, and MGX that intends to invest US$500 billion in US AI infrastructure over four years, with US$100 billion deployed immediately—SoftBank carrying financial responsibility, OpenAI operational. EO 14179 of 23 January made it national policy to 'sustain and enhance America's global AI dominance' and ordered an AI Action Plan within 180 days, delivered on 23 July 2025. But $500 billion is intent, not outlay: by Epoch AI's April 2026 tally the seven US sites totalled over 9 GW of planned capacity while only Abilene was running, at 0.3 GW.
- Related people
- Larry Ellison · Sam Altman
- Related organizations
- OpenAI · Oracle Corporation
- Appears in
- A General History of Information Technology · A History of Cloud Computing
United States District Court for the Northern District of California (via Wikimedia Commons) · Public domain (PD-USGov-Courts) · Commons ↗ Judge William Alsup of the Northern District of California issued the first substantive US ruling on how fair use applies to generative AI training. It split three ways: using books to train an LLM was 'exceedingly transformative' and a fair use; converting lawfully purchased print books into searchable digital copies was also fair use, but for a different reason; downloading more than seven million pirated books from LibGen, PiLiMi and Books3 to build a permanent central library was not. The piracy claims headed for trial. The parties signed a term sheet on 25 August 2025 and a settlement agreement on 5 September for US$1.5 billion, and on 20 July 2026 Judge Araceli Martínez-Olguín granted final approval and entered judgment — 482,460 works on the Works List, about US$3,000 each. The release covers inputs only, and only through 25 August 2025: no output claims and no future conduct are released.
- Related organizations
- Anthropic
- Related terms
- Large Language Model (LLM)
- Appears in
- A General History of Information Technology
Coolcaesar (Wikimedia Commons) · CC BY 4.0 · Commons ↗ OpenAI, founded as a nonprofit in 2015 and fitted with a capped-profit subsidiary in 2019, rebuilt the container. The nonprofit became the OpenAI Foundation; the for-profit became OpenAI Group PBC, a Delaware public benefit corporation. Through a class of stock designated Class N, the Foundation alone appoints and removes the PBC's directors, and at closing it held 26 per cent of the equity — about US$130 billion by OpenAI's own account — plus a warrant paying out further shares if the share price rises more than tenfold over fifteen years. Microsoft holds roughly 27 per cent; the remaining 47 per cent sits with current and former employees and investors. California's attorney general signed a Memorandum of Understanding dated 27 October; Delaware's issued a Statement of No Objection on 28 October — two different instruments, not one. A new Microsoft agreement announced the same day extended the IP licence to 2032 while making any AGI declaration by OpenAI subject to verification by an independent expert panel. The 2019 profit cap was gone, and with it the ceiling on how much capital the company could raise.
- Related people
- Sam Altman · Satya Nadella · Elon Musk
- Related organizations
- OpenAI · Microsoft Corporation
- Related terms
- Artificial General Intelligence (AGI)
- Appears in
- A General History of Information Technology
National Security Agency (Wikimedia Commons) · Public domain (work of the US federal government) · Commons ↗ President Trump signed Executive Order 14409, 'Promoting Advanced Artificial Intelligence Innovation and Security' (91 FR 34565, published 5 June). It gives the Secretary of the Treasury, the Secretary of War through the Director of the NSA, and the Secretary of Homeland Security through the Director of CISA 60 days—to 1 August 2026—to build a classified benchmarking process for the advanced cyber capabilities of AI models, with the NSA Director alone deciding what counts as a 'covered frontier model'. Developer participation is voluntary, and government access runs for up to 30 days before a model's release to other trusted partners—not before public release. Section 3(c) states that nothing in the section may be construed to authorise mandatory licensing, preclearance, or permitting. On every axis—public versus classified, mandatory versus voluntary, statutory versus collaborative—it inverts Biden's EO 14110. As of 12 August 2026 no Federal Register document answering the 1 August deadline had appeared.
Daniel Torok / US Department of Commerce (Wikimedia Commons) · Public domain (work of the US federal government, 17 U.S.C. §105) · Commons ↗ On 12 June 2026 the Bureau of Industry and Security, over Commerce Secretary Howard Lutnick's signature, directed Anthropic by letter to suspend all access to Claude Fable 5 and Claude Mythos 5 by any foreign national anywhere—including the company's own foreign-national staff inside the United States. Anthropic disabled both models for every customer. The trigger was an Amazon research report describing a prompt that bypassed Fable 5's safeguards. The controls were withdrawn on 30 June and Fable 5 returned worldwide on 1 July. The common framing that this was the first time export controls reached model weights does not survive checking: ECCN 4E091 of the January 2025 AI Diffusion Rule already covered closed model weights, and was rescinded in May 2025—and in any case no weights moved here, only access to models running on Anthropic's servers. What was new was export-control authority aimed at a named company's named models, reaching continuously available API access.
- Related organizations
- Anthropic · Amazon
- Appears in
- A General History of Information Technology · A History of Cybersecurity