(Mandatory Credit: Bernie Nunez /Allsport)
Deep Thinking by Gary Kasparov (London: John Murray, 2017) £20
Among the many myths of genius that are routinely indulged is that of the chess grandmaster. We have made these men, women and somewhat precocious children, emblems of higher brainpower by which some super-powered mental circuitry elevates them to a level of pure and abstract thought. When James Bond needs a particularly meaty villain to fight, you can be sure that villain plays chess.
Chess as the pinnacle of human thought is such a popular preconception that it is perhaps unsurprising that early pioneers in Artificial Intelligence believed that solving the mystery of chess might lead to them solve the mystery of human cognition. It led, in just a few decades, to the first victory of a computer over a chess world champion when IBM’s Deep Blue beat Gary Kasparov on the 10th February, 1996.
The story of that victory is the subject of Gary Kasparov’s Deep Thinking yet what holds the book together is something more critical to the wider world. The book is subtitled ‘Where Machine Intelligence Ends and Human Creativity Begins’ and, to use a chess analogy, this is Kasparov’s chance to sacrifice his sport for his faith that the unique quality of being human will never be replicated by a machine. He must deflate the myth of chess in order to make the astute point that “intelligence” is something quite different to competence around the chess board.
In doing so, Kasparov defies his reputation (itself another of those chess myths) of the player with the perpetual scowl to lead us through a story that is light, readable, often funny, but always insightful. It is part confessional and part appeal to the Grand Jury of History. He admits, at the outset, to being human. He was (and, presumably, is) flawed and this lends the story the most intriguing dimension. In defeat, he said and did a few regrettable things. Yet this dichotomy between Kasparov’s human frailty and Deep Blue’s artifice belies the deeper reality of the battle. The book exposes the politics behind a competition that IBM believed would help revive its business (as, indeed, its share prices surged after Kasparov’s defeat). It’s a story that sets Deep Blue’s victory in the context of human psychology, gamesmanship, and, ultimately, the cold practical business of corporate success.
These revelations rightly undercut Big Blue’s achievement and it’s a healthy reminder that the machines, as yet, are passive players in a world still dominated by human motives. Chess, meanwhile, was always a problem that lent itself to computation. It is a problem that can be solved through big numbers. There might only be a finite number of moves in any one position but calculating subsequent moves rapidly generate huge numbers of permutations the “deeper” you traverse the every growing tree of solutions. Humans have never played chess that way but computers can. They can simply “brute force” their way to victory by constantly recalculating their path to the easiest victory. The success of Deep Blue was always a matter of “when” and not “if” a computer would triumph over the best human player.
This is important to realise. A machine that’s extremely good at chess does not, in any sense, threaten humans when it comes to the really important things that we do. It also raises the question as to what, really, is “intelligence”, artificial and otherwise. Why is the chess champion considered a genius and the person who can make beans on toast considered quite ordinary?
On the face of it, the question is ridiculous because we know there’s a difference. For a start, most of us can slap beans on a round of toast but very few of us can play a sharp Sicilian position. Yet, when it comes to a machine replicating human consciousness, it’s the former activity that increasingly lends itself to proper research. AI has made hardly any progress in overcoming these apparently trivial yet deeply complicated tasks. Even in an area where we’ve seen rapid progress, such as speech recognition, the achievements are still relatively small. The logic behind Apple’s Siri or Amazon’s Alexa is really quite primitive compared with true human speech and the levels of cognition that lie beyond it.
Where Kasparov perhaps underplays his position is that he does not get drawn into the wider ethical concerns about AI. Since his retirement in 2005, Kasparov has been increasingly vocal in politics, earning himself a reputation as one of Putin’s fiercest critics. What I really wanted to read was Kasparov’s view of AI as related to our current plight. Kasparov is healthily pragmatic about the use of AI and rightly expresses the optimistic view that we’re a long way from a Terminator style rise of the robots. Certainly, AI offers the world a chance to remove humans from dangerous or dull work. What it doesn’t answer, however, is how the world copes with a new economy in which there are few low skilled jobs. Donald Trump promising his voter base that he’ll bring back coal jobs might not have been realistic but, in terms of psychology, it was all powerful.
Trump in a sense is the factor missing from Kasparov’s calculation in that it’s not the obvious application of AI that should worry us as much as its covert influence. Computers might never replace humans in terms of pure creativity but such optimism overlooks how far even brute force methods are already impacting our lives. If Deep Blue has a legacy, it’s not in the world of chess but in the world of Big Data and the hermeneutic algorithms behind targeted advertising and the bot networks that might well have influenced elections in the past and are sure to influence elections in the future? What, for example, would Kasparov make of the neural networks currently being taught to look at human faces in order to determine a person’s sexual orientation, their political allegiance, and even their IQ? What if Deep Blue could have examined Kasparov’s face during his match and added its own level of psychological analysis to its game play?
That, one would hope, could be a book for the future since Kasparov is perhaps uniquely situated to look into these real concerns. As it is, Deep Thinking is a welcome antidote to all the mythmaking that surrounds AI as well as providing even a lay audience with an insight into the workings of the mind of chess’s most respected champion. Kasparov might well have been beaten by a computer but, really, Big Blue’s victory was hollow. It’s the process and not the result that often defines us as human. In telling his side of the story, Kasparov wins the peace even if he did not win the war. The point, surely, is that only a human could tell this tale and we should be thankful that Kasparov took time to do just that.