Dartmouth College (AI Conference)

Contribution to AI

The name was the deliberate part. Drafting the August 1955 proposal that brought a shifting cast of researchers to Hanover the following summer, McCarthy wanted a label that carried no prior allegiance — not cybernetics, which meant Wiener, and not automata studies, which had already attracted papers about machines rather than about thinking. What he and Minsky, Rochester and Shannon asked the Rockefeller Foundation to fund was a two-month study resting on a single conjecture: that every feature of intelligence could be described precisely enough for a machine to simulate it. The foundation gave them roughly half the $13,500 requested. As a workshop it barely functioned — attendees arrived and left on their own schedules, kept to their own projects, and produced no shared result. Newell and Simon turned up with the one programme that already worked, the Logic Theorist, along with the list-processing ideas that made it possible; Solomonoff spent his weeks there on inductive inference. What survived was the framing document. It listed the problems — natural language, abstraction, self-improvement, neuron nets, randomness and creativity — that the next thirty years of work would divide themselves among, and it attached a term that stuck to all of them. A summer of poor attendance gave a scattered set of interests one name and one agenda.

Drafted with AI and edited by hand (claude-opus-5, reviewed 2026-08).