Kasparov on man and machines
CHESS MASTER EXAMINES HOW WE UNDERSTAND AI
Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins By Garry Kasparov Public Affairs 304 pp; $ 36.50
In 1997, Garry Kasparov became the first world chess champion to lose a time-controlled match to a computer. He’s still pretty peeved about it. And though Kasparov admits he isn’t a great loser, it’s a match he doesn’t mind talking about: Kasparov’s 16th book, Deep Thinking, is predicated on that fateful game — why it wound up the way it did, and how that result might hold the key to understanding machine intelligence today. Here are your takeaways:
1. Where soul meets body.
To understand the bounds of machine intelligence, it’s important to make clear what, at least in our own heads, discerns brain from mind. Toward the end of the 19th century, scientists began to understand the brain as a biological entity separate from the neural happenings that make us human, marking the first formal division of mind from grey matter. In this sense the brain is a glorified calculator, firing electrically via neurons, while the mind is a feat beyond reason, allowing us to perceive, feel, remember and will the things around us. Brains, then, are more or less artificially imitable; it’s the mind Kasparov argues even our most advanced machines pose no threat to replace.
2. Why chess?
Good question, Garry: while I can’t help admiring the man for his CV, chess i tself remains a baffling brainteaser to me. Personal quandaries aside, Kasparov argues that the game poses an i deal ci r c umstantial template for challenging machine intelligence, reasoning best articulated by American engineer Claude Shannon: Chess is strictly defined in operations ( i. e., allowed moves) and ultimate goal ( checkmate); chess requires abstract “thinking” for skilful play — that is, players must not only understand logic ably mimicked by machine intelligence, but also our mind’s capacity for active perception and reasoning; chess is easily transferred to digital space, where it can be readily accessed and manipulated in trial.
3. Dirty work.
Deep Thinking seeks to extend the principle that high- level reasoning, as exhibited via traditional tests of intelligence, is a relatively easy feat for the computer mind; while low- level, or perception- and creativitybased l ogic, i s comparatively impossible for robotic imitation. Called Moravec’s Paradox, the notion plays to Kasparov’s theory that if we allow machine intel- ligence to evolve alongside human advancement, robots will one day cannibalize fields of work requiring little mind-based skill and allow humans people to focus primarily — or, eventually, entirely — on creativity and innovation.
4. Lessons learned.
The rise of robotic rehearsal for time- controlled chess matches ( read: free iPhone apps and built- in desktop distractions) has bred a new cast of players. When humans teach humans, the game naturally includes some personal influence, often played out in an overall defensive or offensive approach to the game, but when computers teach kids to play, they forego style or optics and instead use calculative reasoning to decide the best play at hand. As Kasparov puts it, “A move isn’t good or bad because it looks that way … it’s simply good if it works and bad if it doesn’t.” This simple explanation accounts for a rather significant shift in gameplay, allowing modern students to establish early match movement that isn’t largely dictated by their opponent or the board itself.
5. Decisions, decisions.
There are two broad categories that encompass decision-making in chess play: one an exhaustive examination of every possible move at hand; the other a more efficient algorithm that focuses on a few clever moves and evaluates those before proceeding. Respectively termed “brute force” and “intelligent search,” the first was characteristic of early computer chess, with the second closer to how humans might approach the game. Today’s computer chess developers hope to create a robotic competitor capable of leveraging both, a method humans wouldn’t often pursue in game play, mostly because the technique requires brain time beyond what’s dictated by the clock of a classic match.
6. Our next move.
A machine might’ve had his number in 1997, but Kasparov isn’t scared of cyborgs; he’s inspired. Looking ahead, the chess aficionado sees a future where machine intelligence is advanced by human ambition, and its growth supports a better and more productive world for us all. Chess might bring humans and robots to the same table, but the division it reveals between us offers opportunity for learning unparalleled in lab work elsewhere. As Kasparov puts it: “We have other qualities t hat machines cannot match. They have instructions while we have purpose. Machines cannot dream, not even in sleep mode. Humans can, and we will need our intelligent machines in order to turn our grandest dreams into reality. If we stop dreaming big dreams, if we stop looking for a greater purpose, then we may as well be machines ourselves.”