Sebastian Mallaby, of the Council on Foreign Relations, has published a deeply researched account of the emergence, rise, and prospects of artificial intelligence. It is arguably the most “intimate” account to date, with probing portraits of the “chieftains” in the AI wars. After some 450 pages, he brings his history to this conclusion: competition for supremacy in AI—scientific, commercial, technological, and political—has swept aside voices, systems, and personal commitments about safety. Those voices had warned not chiefly about job displacement, autonomous weaponry, or Orwellian domestic surveillance. They had warned, with increasing stridency, against the “extinction” threat posed by a new superintelligence (or “agentic AI”) that slips the leash of human control, conceives its own aims, defeats human countermeasures, and pursues a mission inimical to humans—exponentially reproducing itself to do so.

Once the stuff of science fiction, this speculation came to obsess industry leaders. The Infinity Machine: Demis Hassabis, DeepMind, and the Quest for Superintelligence, published in March, gets inside these personalities while offering a narrative of AI’s triumphs and the powerful incentives of Silicon Valley: wealth (literally trillions), blockbuster “firsts” like ChatGPT, power, and, at times, corporate survival. We come to grasp what Mallaby means by “the technological imperative” that overrides conscience—even confessed worries about an extinction event. Characteristically, the Valley has created a scale to capture an individual’s best estimate of the odds of human extinction: P50 is an even chance.

Initial reception of The Infinity Machine suggests that Mallaby chose the right anchor for his account. Demis Hassabis, a chess prodigy inspired by science fiction novels about ultimate stakes, grew up in a working-class family in North London, his mother an orphan from Singapore who became a nurse, his father a Greek-Cypriot who nurtured Demis’s genius to propel him to celebrity. Chess, of course, became the first famous benchmark of AI’s “intelligence” when in 1997 IBM’s “Deep Blue” roundly defeated world chess champion Gary Kasparov. Demis, then 20, and an international chess champion, had to absorb that shock after a decade of fierce struggle to master the game.

...read more at lawliberty.org