In 1955, John McCarthy, a young mathematics instructor at Dartmouth College, wrote a proposal to the Rockefeller Foundation. The title contained a new phrase: "Artificial Intelligence on a Summer Research Project." McCarthy was trying to name something that existing labels — "automata theory," "cybernetics" — refused to cover. The term has stuck not because it failed but because it caught on.

The workshop that ran six to eight weeks on the Dartmouth campus in summer 1956 brought together the leading thinkers of the moment: McCarthy (who would create LISP), Marvin Minsky (working on neural networks at MIT), Claude Shannon (founder of information theory), Nathaniel Rochester (architect of the IBM 701), Allen Newell and Herbert Simon (creators of the Logic Theorist program), Arthur Samuel (the learning checkers program), and others. The proposal contained a confident hypothesis: "every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it."

The workshop did not produce a single tangible result; it was more a definition of a field. Newell and Simon's Logic Theorist — a program that proved mathematical theorems — was at the centre of debate that summer. McCarthy published LISP in 1958, which would become the fundamental language of AI. Minsky founded the MIT AI Laboratory in 1959. Simon claimed in 1965 that "machines will be capable, within twenty years, of doing any work a man can do" — a prediction that did not come true.

The history of AI has been a cycle of booms and busts: two "AI winters" in the 1970s and 80s as funding dried up; the field reheated by deep learning from the 2010s; ChatGPT in 2022 took the field fully into public consciousness. The term McCarthy used in 1955 has become, 75 years later, the name of the largest technology-investment wave in the world — but the fundamental Dartmouth question remains open: can "general intelligence" actually be imitated?