About this timeline
What it is
The History of AI follows artificial intelligence from 1763 to the present, from the disciplines that made it possible to today's race toward AGI.
Its central idea is that AI did not grow out of a single line of work. It grew out of five, which advance separately for almost two centuries, draw close in 1943 and in 2012, and have run as one since 2022. That is why it is drawn as five streams rather than as a list of dates.
How to read it
Each dot is a milestone: a dated event with the people behind it and its consequences. Four things place every milestone.
- The era. There are five periods: The foundations (1763–1942), AI is born (1943–1979), From the expert boom to the statistical turn (1980–2011), Deep learning and foundation models (2012–2021), and Generative AI, reasoning, and agents (since 2022).
- The field. Within each era, the area of work it belongs to, such as symbolic AI or neural networks.
- The lineage. The deeper current that feeds it and gives it its color: symbols and logic, data and learning, neurons and networks, machines and computation, or acting in the world.
- The type. What kind of event it is: a theory (an idea or a theoretical result), a publication (a paper, a book, or a report), an implementation (a working system or model), a popularization (the moment an existing idea spreads), an application (a use with effects outside the lab), or an institutional event (foundings, funding, regulation, prizes, and meetings).
The links between milestones show what led to what, and they tell a documented direct influence apart from a looser inheritance, an independent convergence, or a mere coincidence in time. Threads chain several links together to follow an idea from start to finish.
In the scene, time does not move at the same pace across the whole timeline: each era takes up space in proportion to what happened in it, not to how many years it lasts. That is why the first 180 years take up less room than the last five.
How it is made
Each milestone is documented with its original source (the paper, book, patent, or official announcement) and with histories of AI that place it in context. Dates are those of first publication or first public presentation.
The timeline's full list of sources is in the Index. The sources for each milestone are on its entry.
If you find a mistake
A reference work gets better through its readers. If a date, a name, or an attribution looks wrong to you, write to thehistoryofaiweb [at] gmail.com, naming the milestone and, if you can, the source that contradicts it.
Your data
This website requires no registration, shows no advertising, and uses no analytics, advertising, or tracking cookies. What your browser stores and how visits are counted is explained in the legal notice.
Last updated: November 15, 2026