The founder of the Future of Life Institute explores one of the most intriguing scientific frontiers, artificial general intelligence, and how humans can grow along with it.
Nowadays, computers read, learn, recognize faces, translate languages, and consult other computers. They don’t yet think, but the contingent of researchers who believe that they will never be smarter than humans is steadily shrinking. In this expert but often wildly speculative rumination, Tegmark (Physics/MIT; Our Mathematical Universe: My Quest for the Ultimate Nature of Reality, 2014, etc.) joins the fierce debate on what will happen when AGI reaches human level and beyond. He dismisses tabloid scenarios of rampaging robots but warns, “we might create societies that flourish like never before…or a Kafkasque global surveillance state so powerful that it could never be toppled.” The author defines intelligence as the ability to accomplish complex goals. This sounds trivial until he points out that both brains and computers are able to do this. Since computers are improving faster than brains, superhuman AGI will happen, and a beneficial outcome is not guaranteed. Thus, autonomous, self-driving cars will save lives. Autonomous battlefield drones will save soldiers’ lives, but keeping them away from rogue nations, terrorists, and criminals will prove impossible. In the early chapters, Tegmark portrays near futures that range from Utopian to Orwellian. Later in the book, he delivers a vision of the far future: a universe filled with the products of superintelligence, with organic Homo sapiens a distant memory. Throughout, the author lays out his ideas in precisely detailed scenarios. Many read like science fiction; others, such as a fine chapter on the nature of consciousness, are simply good popular science.
Prophesies have a dreadful record, but they are also endlessly fascinating. Readers may balk now and then—Tegmark’s solutions to inevitable mass unemployment are a stretch—but most will find the narrative irresistible.