When was AI invented?
AI was invented in 1956, at the Dartmouth Summer Research Project — the workshop that gave the field its name and its research agenda. The name itself was written down earlier, in a proposal dated August 31, 1955. Alan Turing framed the question in 1950. As of 2026, AI is 70 years old.
Why — the first-principles explanation
The date you pick depends on what you count as "AI." Three dates have real claims, and they are only a few years apart.
1950 — the idea gets a testable form. Alan Turing published "Computing Machinery and Intelligence" in the journal Mind. He argued that "Can machines think?" is too vague to answer, and replaced it with the imitation game: put a human judge at a keyboard, let them chat with a machine and a person, and see if the judge can tell which is which. That move — turning a philosophy question into an engineering benchmark — is what made AI a science instead of a debate. 1955 — the field gets a name. John McCarthy, with Marvin Minsky, Nathaniel Rochester, and Claude Shannon, wrote a proposal dated August 31, 1955 asking the Rockefeller Foundation for $13,500 to run a summer study. That document is where the phrase "artificial intelligence" first appears in print. McCarthy later said he picked the name partly to avoid being lumped in with cybernetics. 1956 — the field actually starts. The workshop ran at Dartmouth College in Hanover, New Hampshire over roughly eight weeks that summer. It is the date almost every textbook uses, because it is when the researchers showed up, argued, and left with a shared label. Newell, Simon, and Shaw brought the Logic Theorist, a program that proved mathematical theorems — the first program most historians call AI.
Why the confusion? Because "invented" fits products, not fields. Nobody asks when chemistry was invented. AI got a founding event with a date, so people treat it like a gadget with a patent. If you need one number, say 1956 — that is the field's official birthday. If you want the honest version, say the idea is from 1950, the name is from 1955, and the field is from 1956.
An example that makes it click
Imagine asking when a restaurant was invented. Someone had the idea in the shower in 2020. They wrote the business plan and picked the name in 2021. The doors opened in 2022. Which year did the restaurant get invented?
AI works the same way. Shower thought: Turing, 1950. Business plan and name: McCarthy's proposal, August 1955. Doors open: Dartmouth, summer 1956. All three answers are defensible, but the one everyone agrees to write on the sign is 1956 — because that is when people actually walked in.
Key facts
- The Dartmouth Summer Research Project on Artificial Intelligence ran in summer 1956 and is widely treated as the founding event of the AI field.
- The proposal that first put "artificial intelligence" in print is dated August 31, 1955.
- That proposal requested $13,500 from the Rockefeller Foundation for a two-month, ten-person study.
- Alan Turing's foundational paper appeared in Mind, volume 59, pages 433-460, in 1950 — six years before the term existed.
- Warren McCulloch and Walter Pitts published a mathematical model of the neuron in 1943, the ancestor of today's neural networks.
- Counting from 1956, AI turns 70 years old in 2026.
▶ The 60-second explainer (script)
When was AI invented? Short answer: 1956. Longer answer: it depends what you mean, and there are three dates worth knowing. Nineteen fifty: Alan Turing publishes a paper in the journal Mind asking "Can machines think?" He decides that question is too fuzzy, so he invents a test instead — a judge types to a human and a machine and has to guess which is which. That's the moment the idea became something you could actually build toward. Nineteen fifty-five, August thirty-first: John McCarthy writes a funding proposal with Minsky, Rochester, and Shannon. They want thirteen thousand five hundred dollars for a summer of work. That paper is where the words "artificial intelligence" appear in print for the first time. Nineteen fifty-six: the workshop happens at Dartmouth College in New Hampshire, about ten people, roughly eight weeks. Everybody leaves calling their work AI. That's the birthday textbooks use. So the idea is 1950, the name is 1955, the field is 1956. And in 2026, that makes AI seventy years old — older than the microchip in your phone, older than the internet, and a lot older than most people assume.
What authoritative sources say
People also ask
Was AI invented in 1956 or 1950?
Both dates are used. 1956 is the field's founding workshop at Dartmouth and the standard textbook answer. 1950 is when Turing published the paper that framed the whole question, six years before anyone said "artificial intelligence."
Who invented AI?
John McCarthy coined the term in 1955 and organized the 1956 workshop, so he gets the most credit. Marvin Minsky, Nathaniel Rochester, Claude Shannon, and Alan Turing are also routinely named as founders.
Why does AI feel new if it was invented in 1956?
The ideas are old but the hardware and data are new. AI went through two long funding droughts, called AI winters, in roughly 1974-1980 and 1987-1993. What changed recently is cheap parallel computing, huge text datasets, and the transformer architecture from 2017.
Was AI invented before computers?
No, but barely after. The first programmable electronic computers date from the mid-1940s. Turing's paper came in 1950, before most people had ever seen a computer, and the field was named in 1955.
The same question, asked other ways
- When was AI created?9,900/mo
- When did AI come out?4,400/mo
- How long has AI been around?4,400/mo
- When did AI start?2,900/mo
- When was AI made?2,400/mo
- How old is AI?1,900/mo
- When was AI first created?1,000/mo
- When was the first AI created?720/mo