Tracing the Origins of a Forgotten Narrative

When you ask an artificial intelligence professional to date the birth of their field, the answer is almost unanimous: 1956, the Dartmouth Conference. It was there, at this American university, that John McCarthy officially coined the term "artificial intelligence." A designation that was as much a marketing move to attract funding as it was a scientific revolution.

Go back a few years, and you'll find a second point of origin: Alan Turing's foundational 1950 paper, which poses the question "Can machines think?" and introduces the famous test that bears his name.

Two dates, two names, two Western references. This is the dominant narrative of AI history.

Beyond the Dominant Narrative

Yet, if we accept a definition of artificial intelligence not as an academic field born in 1956, but as humanity's millennia-old attempt to encode thought into matter, then history changes radically. The automata of ancient Greece, African binary systems, medieval Islamic robotics: all chapters erased from the grand narrative.

This observation gave birth to Project Avalon, which became History-AI: an attempt to tell the story of artificial intelligence not over 75 years, but over more than two millennia. And to do so in French first (then English), for a curious but non-specialist audience.

The Methodological Question

But how do you produce such a project in just a few days, with the necessary scientific rigor and demanding editorial quality? This is the methodological question this note proposes to explore.