The Dartmouth Workshop of 1956

In the AI Story

The Dartmouth Workshop sits at the origin of every subsequent conversation about AI. When contemporary writers ask "what is artificial intelligence?", the historical answer runs back to this summer in Hanover, New Hampshire. The proposal's famous phrasing — that "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" — set the field's animating ambition and its founding controversy.

The proposal's optimism was characteristic of its moment. McCarthy and Minsky expected significant progress in a single summer of concentrated effort. The under-delivery of that expectation became a recurring pattern in AI history, eventually producing the term "AI winter" for periods when funding and enthusiasm collapsed. The current era — the large language model revolution — is the first period since Dartmouth in which AI has exceeded its most ambitious founding claims rather than fallen short.

In the decades after Dartmouth the workshop's name became a shorthand for AI's entire founding ambition, rather than a specific eight-week event. This is historically slippery: Dartmouth was a modest academic meeting whose main accomplishment was establishing the research community. The grand claims of the 1955 proposal — that language, abstract thinking, and self-improvement could be substantially addressed in a summer — have become embarrassing in retrospect, but they are no more ambitious than claims made routinely in today's AI labs. Reading the proposal now is partly a humbling exercise in how timelines for ambitious research programs actually play out.

Origin

The workshop was proposed in August 1955 in a funding application to the Rockefeller Foundation, co-authored by John McCarthy (Dartmouth), Marvin Minsky (Harvard), Nathaniel Rochester (IBM), and Claude Shannon (Bell Labs). The Rockefeller Foundation awarded $7,500, and the workshop ran from June 18 to mid-August 1956 at Dartmouth College.

Attendees included Allen Newell and Herbert A. Simon (whose Logic Theorist was demonstrated there — the first running AI program), Ray Solomonoff, Arthur Samuel, and Oliver Selfridge, among others.

Key Ideas

The naming. McCarthy proposed "artificial intelligence" partly to distinguish the new field from cybernetics (Norbert Wiener's term) and from "automata theory". The choice proved durable.

The ambition. The proposal asserted that machine learning, natural language, neural networks, and abstract reasoning could all be substantially advanced "in a summer". This optimism set the tone for early AI.

The distance between communities. The symbolic-AI tradition (McCarthy, Minsky) and the connectionist tradition (neural networks, represented at Dartmouth by discussions of perceptrons) separated soon after and did not fully reunite until the deep-learning era.

The event as institutional birth. Dartmouth matters less for what was produced there than for the research community it defined.

Who was absent matters. Dartmouth was almost entirely white and male, and the connectionist tradition that produced neural networks was underrepresented. The intellectual narrowness of the founding workshop helped determine which approaches the field took seriously and which it sidelined for decades — a lesson echoed in contemporary debates about AI's social composition.