Genesis Artificial intelligence, hope, and the human spirit

The late former secretary of state Kissinger and former Google CEO Eric Schmidt, who previously collaborated on 2021's The Age of A.I., team up with technology consultant Craig Mundie for this warmed-over consideration of how AI might change the world. Their prognostication careens between the dystopian and utopian. On the one hand, they caution that countries might weaponize AI to sic supercharged computer viruses on their enemies' digital infrastructure, and that humans' inability to understand how AI reaches its conclusions might "catalyze a return to a premodern acceptance of unexplained authority." On the other hand, the technology might raise living standards by devising cheap "synthetic substitutes" for in-demand physical resources like oil and gas, or extend lifespans by editing genomes. Unfortunately, the authors offer precious little in the way of evidence and lean heavily on speculation. For instance, their assertion that machines could one day achieve sentience is grounded only in their faith in the inevitability of technological progress. They give short shrift to AI's well-documented limitations, and the policy recommendation to pursue AI development in a manner consistent with humanity's "moral convictions" is a vague cop-out. This doesn't add anything of significance to the conversation on AI. Agent: Andrew Wylie, Wylie Agency. (Nov.)

Unsettling thoughts on AI from some big-league thinkers. The book opens with a worshipful tribute to Kissinger, who died last year at age 100. He apparently thought deeply about this subject and "closely mentored" his two collaborators on the "diplomatic alignment of humans" in their relationship to AI. Kissinger was a prominent Secretary of State and Nobel Peace Prize winner, Eric Schmidt was the chairman and CEO of Google, and Mundie was a chief research and strategy officer for Microsoft. Recognized AI experts, the latter two work hard to avoid repeating themselves. Within the past decade, specialists who have spent their lives warning against comparing a computer to a human brain have changed their minds. AI engineers now agree that they are trying to build something modeled on and superior to the brain without fully understanding it. Human brains are limited by the size of the human skull, but AI can grow without limit. The average AI supercomputer is already 120 million times faster than the human brain. Unlike ordinary computers, its mapping of the world is not programmed but learned. Perhaps in a nod to Kissinger, the book delves heavily into the role of AI in government. Its value lies in its potentially perfect knowledge, but that is a two-edged sword. Faced with a machine that always makes the correct choice, humans and their leaders may object to surrendering their free will, but autocrats may perk up. Readers may scratch their heads at the authors' conviction that AI builders assume their creation will lead to a golden age of abundance, eliminating poverty, inequality, and the necessity of work. A long section devoted to problems of a life devoted exclusively to leisure contains ingenious solutions, but it's not a subject that provokes great controversy. Astute if often oddball insights. Copyright (c) Kirkus Reviews, used with permission.