ARTIFICIAL INTELLIGENCE

McPherson presents the evolution of artificial intelligence—machines with the “humanlike ability to reason and solve problems.”

That definition opens McPherson’s tour d’horizon of artificial intelligence, immediately placing readers on shaky ground. Philosophers have been debating “to reason” since long before Descartes. There is little doubt that McPherson richly explores the women and men who develop machines to do the drudge work of mechanical production and everyday life, but do either the amusingly crafty Watson, which took down the Jeopardy! game show champs, or Deep Blue, which humiliated Garry Kasparov, qualify as “a truly thinking machine, able to learn on its own and modify its own programming without human input”? The ability for a machine to reckon if/then is part of its programming. Sentience, which includes feeling, is stickier. How is it possible, as McPherson writes, that a machine programmed by humans “might not share human social and ethical values—such as notions of fairness, justice, and right and wrong”? Throughout, there’s too much supposition and not enough science; emblematic of this is a failure to convey exactly how Google Brain arrived at the concept of a cat without being commanded to: “All on its own, it had developed the concept of ‘cat.’ ”

McPherson conveys the thrill of the possibility inherent in AI, but she’s frequently a giant step ahead of the game.