In the future, will we all be cyborgs connected to computers?
Garry Kasparov explores how people, not artificial intelligence, are a threat to humankind
If the startling progress of artificial intelligence suggests we’ll soon be bowing down to our robot overlords, Garry Kasparov advises us: Don’t panic.
Yes, that Garry Kasparov, the Russian former world chess champion who famously lost to IBM’s computer Deep Blue in 1997, sparking existential dread that humans were on their way to obsolescence.
But as Kasparov reminds us in his new book, Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins, he had defeated Deep Blue the year before and, after his loss, IBM refused to offer a rubber match.
On the phone from his New York City home, Kasparov maintains that this was “a pure corporate decision: they evaluated that the next match would be uncertain.”
Looking back at that time, Kasparov realizes how much help the computer had behind the scenes, from fallible humans.
And even today’s much-improved chess programs, it seems, lose out to humans and machines working in tandem. In other words: if you can’t beat IBM, join ’em.
Human-machine teams can take advantage of the former’s intuition and the latter’s deduction, Kasparov says, and they point the way to an ideal collaboration in our work and our daily lives. He calls Tesla guru Elon Musk’s latest venture, the development of a brain/computer interface, “a move in the right direction.”
In the future, will we all be cyborgs with chess engines connected to our brains?
“We don’t know,” Kasparov says. “And many people are afraid of hearing, ‘We don’t know.’ I encourage it. If you don’t know, that’s not the reason to stop.”
Since retiring from professional chess in 2005, Kasparov has become a polymath: campaigning for human rights, organizing opposition to Vladimir Putin, giving corporate speeches and spending time at Oxford University’s Future of Humanity Institute, where researchers have worked toward “whole brain emulation” — or recreating the work of a brain in software.
Kasparov’s book, which was out Tuesday, offers a potted history of AI, focusing on the quest to produce an unbeatable chess-playing machine, which became a common goal among researchers because of what he calls the questionable “mystique” that chess prowess represents intelligence.
In the early days, programmers wanted machines to analyze patterns and zoom in on potentially good moves (as humans do), but with the advance of calculating power it became much more effective to try brute force: figuring out potential outcomes of all possible moves in one particular position, a certain number of moves ahead.
This worked tremendously for chess but not for producing smarts: Deep Blue, Kasparov insists, was in fact “as intelligent as an alarm clock.” Having beaten Kasparov, IBM took its AI in more promising directions — toward programs that gather data and then extrapolate.
Other companies have done so too: a version of Google’s self-learning computer program Alpha Go has beaten many of the world’s best players of Go, the ancient Chinese strategy game that is far more complex than chess.
There is, however, a troubling technical issue.
The program, Kasparov says, learns in ways “that are impossible for us even to contemplate . . . The creators don’t know why one version of Alpha Go plays better than the other. If there’s a glitch in Alpha Go, I’m not sure they know where to look.”
An unfathomably wonky game-playing program is one thing, but what about AI-based stock traders, or doctors, or selfdriving cars, where glitches could be disastrous? Humans can “coach” machines and, in Kasparov’s view, our guidance and vision will enable both machines and humans to work better.
And we might as well make the most of this possibility, because the march of these programs, and their encroachment on our employment, is unstoppable.
“It’s progress. I think the only answer is to move forward, to create new industries, new opportunities where machines will have to rely on our intuition since we’ll be entering new zones — like reactivating space exploration, going deep underwater, creating situations where our creativity prevents us from becoming redundant.”
But can everyone participate? His vision sounds suspiciously utopian. What do we do, for instance, about veteran industrial workers who are being laid off, for whom it’s a stretch to retrain as software designers? Kasparov admits he has no ready solution. “They will be hijacked by populists from left or right . . . It is a challenge, but if you try to stop this process by imposing restrictions there’s no way you can win.
“We just have to accommodate ourselves. This is part of our history as the human race. Some people could be more successful, some less. I lost to a machine, but I’m promoting an idea of our cooperation, because that’s a way to move into the future. And I don’t want people to think about AI and machines as surrounded by the fog of secrecy.”
Yet secrecy seems inevitable with AI, as it’s being developed by Silicon Valley. Are we to trust that big companies’ proprietary technology will have our interests at heart? Kasparov admits to some misgivings, but he notes that asking these companies to lift the veils from the way they develop their software could also cause problems. Their work could be then open to abuse from regimes such as Putin’s, looking to adapt them for nefarious ends.
In other words, even though the machines aren’t out to get us, humans in control of these machines may yet be.
“Let’s move into the future with our eyes open,” Kasparov warns. “I hope that my book will help.”
“Some people could be more successful, some less. I lost to a machine, but I’m promoting an idea of our co-operation . . . ” GARRY KASPAROV FORMER WORLD CHESS CHAMPION