The phrase was not coined at a conference. It was coined in a funding application for one, dated 31 August 1955 — nine and a half months before anyone met.

The document is A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence, signed by J. McCarthy (Dartmouth), M. L. Minsky (Harvard), N. Rochester (IBM) and C. E. Shannon (Bell Labs). It opens with a sentence that has aged into comedy: We propose that a 2 month, 10 man study of artificial intelligence be carried out during the summer of 1956 at Dartmouth College in Hanover, New Hampshire.

Two months. Ten people. Seventy years later, still going.

McCarthy is credited with the phrase, and by his own later account it was chosen partly for what it wasn't — a name that kept the new field clear of cybernetics, the established discipline next door. By his own account he chose it to escape that association, and to avoid having either to accept Norbert Wiener as a guru or to argue with him. Which is worth sitting with: the term everyone argues about the boundaries of was picked to stay off somebody else's, in a document whose job was to get money.

But the founders were not vague about the claim. The proposal's central sentence is still the field's whole bet:

The study is to proceed on the basis of the conjecture 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.

Read it carefully, because everything since is a footnote to it. It is a conjecture — explicitly labelled as one. It says in principle, not in practice. And it makes an enormous prior commitment: that intelligence is the kind of thing that can be precisely described at all.

Every serious argument about AI's limits — about understanding, about consciousness, about whether the current systems are doing the thing or performing it — is an argument about that one sentence. Nobody has proven it. Nobody has refuted it.

It is still, seventy years on, a conjecture in a grant proposal.