When researchers debate the origins of artificial intelligence, they typically point to the Dartmouth Conference of 1956 or Alan Turing's foundational 1950 paper. But a growing body of scholarship argues that the true prehistory of AI stretches back far further — to the moment our ancestors first deliberately shaped a rock into a cutting edge.
The cognitive leap required to craft a hand axe — imagining a future form hidden within raw stone and executing a plan to reveal it — represents an early form of the same abstract reasoning that underlies modern machine intelligence. Fire extended this trajectory, freeing early humans from the constraints of daylight and raw nutrition, and creating the surplus time and social bonding that allowed complex thought to flourish.
Language, however, may be the most direct ancestor of today's large language models. The ability to encode knowledge into transferable symbols — spoken, and eventually written — meant that human intelligence no longer died with individual minds. It could accumulate, compound, and be queried by others. In a meaningful sense, every library ever built was a primitive attempt to do what ChatGPT and its contemporaries now do at scale and speed.
This long-arc view of AI development matters for more than philosophical reasons. It reminds us that intelligence augmentation has always been humanity's core survival strategy. The printing press, the calculator, and the spreadsheet were all, in their time, accused of threatening human cognitive labor — just as neural networks are today. Each disruption ultimately reshaped, rather than erased, the human role in knowledge work.
Understanding AI as the latest chapter in a multi-million-year story of tool-making reframes the urgency of current debates. The question has never been whether humans will build cognitive prosthetics. We always have. The question — as it was around the first campfire — is who controls the tools, and to what ends.