The Quest for Artificial Intelligence—and Where It's Taking Us Next
Starred review from January 1, 2017
A history of artificial intelligence and look at the "dazzling (near) future, the changes that lurk just around the corner, and how they will transform our lives forever."During the 1960s, AI seemed to be coming "out of cinemas and paperback novels and into reality," and then the tide receded. Now it's everywhere, in our iPhones, TVs, cars, and even refrigerators. It's a marvelous story, and technology journalist Dormehl (The Formula: How Algorithms Solve All Our Problems...and Create More, 2015, etc.) does it justice. After World War II, when computers began calculating thousands and then millions of times faster than a human, enthusiasts predicted talking robots in a few decades. The author dubs this the era of "Good Old-fashioned AI." Sadly, brute-force calculating enabled a computer to play chess brilliantly, but it couldn't recognize a face, something every 2-month-old baby does. As Steven Pinker said, "the main lesson of the first thirty-five years of AI research is that the hard problems are easy and the easy problems are hard." By the 1980s, funding and media interest had shrunk, but younger scientists turned their attention from programming knowledge one piece at a time to systems that imitate the brain. These "neural networks" employ probability, feedback, potentiation, and inhibition to make sense of data. It works. Computers can't yet think, but they can learn. Google, founded in 1998, was one consequence. The powerful computer named Watson, which easily defeated the best Jeopardy contestants in 2011, succeeded by using analogy and trial and error, not massive stores of facts. This was "deep learning." Computers now recognize faces and the printed word, translate languages, consult other computers, and gather so much information that they can predict our behavior. Not everyone approves. Dormehl lets critics have their say but makes a convincing, often disturbing, but always-entertaining case that that we're in for a wild ride.
COPYRIGHT(2017) Kirkus Reviews, ALL RIGHTS RESERVED.
February 15, 2017
A few years ago, the question, "Can machines think?" would have been answered simply, no. Even the chess-playing program that could beat the best human player was only following a systematic evaluation of all possible positions based upon criteria inputted by its developers. It could not improve beyond that level or learn from experience, nor did it understand the intent of each play or even that it was engaged in a game. However, in this thought-provoking, clearly written book, journalist and filmmaker Dormehl (The Apple Revolution) describes how new programming techniques, such as the use of "evolutionary algorithms," allow a computer to develop its own strategies, improve through practice, come up with original ideas, and even find surprising solutions to unusual problems--that is, to become creative. The author also highlights some of the risks involved in our reliance on these machines, incorporating pertinent examples, including the stock market crash that was caused by machine trading. As we use more and more smart devices, linked through the Internet, these dangers are multiplied. Indeed, many people now worry about the singularity, or the point at which computers can design even smarter computers, which may attain self-awareness and go beyond our ability to control them. VERDICT For anyone interested in developments in the field of artificial intelligence, and a worthwhile acquisition for most libraries.--Harold D. Shane, Mathematics Emeritus, Baruch Coll. Lib., CUNY
Copyright 2017 Library Journal, LLC Used with permission.