Garry Kasparov On AI & Cybersecurity
Garry Kasparov was one of the first, and arguably the most brilliant, people to experience what an increasing number believe will happen to us all in the future: being outsmarted by a machine.
“Chess,” he explains, “was viewed as potentially the ultimate test for a machine's intelligence from the dawn of computer science, by Alan Turing and Claude Shannon and Norbert Weiner.” Kasparov first played against computers in the 1970s, but describes the machines in those days as “laughably weak”.
Even of Deep Thought, the precursor to Deep Blue that he faced in 1989, Kasparov says: “I won quite easily. It was primitive, but if you look back… machines were already making enough progress then for us to realise that it was basically a matter of time.”
By the mid-90s, Moore’s Law had held true for three decades. As in so many areas, the machines appeared to be little more than a novelty until, following the curve of exponential growth, their power became suddenly apparent. “The whole idea that if we had enough time, we would avoid making mistakes,” says Kasparov, “was ignorant. Humans are poised to make mistakes, even the best humans. And the whole story of human-machine competition is that the machines, first it's impossible [that they could play], then the machines are laughably weak, then they are competing, for a brief time, and then, forever after, they are superior.”
But the inevitability of the machines’ success, says Kasparov, is not a matter of brute force, but of reliability. “Machines have a steady hand. It's not that machines can solve the game”, the number of possible moves is so high that, even calculating at 200 million moves per second, it would have taken Deep Blue longer than the life of its opponent, or the solar system or quite possibly the universe itself, to calculate them all, “it's about making moves that are of a higher average quality than humans.”
The machine, says Kasparov, need never fear losing its concentration because it can never feel fear and it has no concentration to lose. “It doesn't bother about making a mistake in the previous move. Humans are by definition emotional. Even the top experts, whether it be in chess, or video games, or science - we are prisoners of our emotions. That makes us easy prey for machines, in a closed system.”
In 1997, Kasparov played his second match (he had won the first) against the IBM supercomputer Deep Blue and lost in the deciding game. He had been the World Champion since 1985, and would remain the world’s highest-rated human player until his retirement in 2005. He found losing to a machine to be “a shocking experience,” although this was partly, of course, because “I haven't lost many games… Now, two decades later, I realise it was a natural process.”
But Kasparov does not think humans are about to be replaced entirely by machines. Even in cybersecurity, where automation and machine learning are necessary, “It's not a closed system, because there are no written rules.
Actually, it's one of the areas where human-machine collaboration will have a decisive effect. I think it's naïve to assume that machines could be totally dominant, because the angle of attack can change. There are so many things that can change. It's an unlimited combination of patterns that can be manipulated.”
He is optimistic about technological development because “I read history books, and I know that it's happened before. It's a cycle. New technology always destroys jobs and industries, before it creates new ones. So if we try to slow down AI development, it will still kill jobs, but it will slow down the creation of new industries. It will prevent us from generating enough income to think about the people who have been left behind.”
“We can learn from history that there were negative effects when automation killed millions of manufacturing jobs; 100 years ago, 30 per cent of the American population worked in agriculture. Now it's two per cent. The only difference is, that today, machines are going after people with degrees, political influence and Twitter accounts. So now it's a big story. But it's a natural process; each industry must feel the pressure of innovation.”
For Kasparov, the need for governments to embrace technological change is more than simply an economic issue. As the chairman of the Human Rights Foundation he works with dissidents around the world to promote human rights and liberal democracy, and as a Russian he composes strategy against a singularly dangerous opponent: Vladimir Putin. The antidote to the Russian president’s powerful campaigns of espionage and disinformation, he says, is a combination of computing power and human intuition.
“The fake news industry isn't about selling you a hard product. It's not about communism, or fascism. It's selling you doubt. That's why I call Putin a 'merchant of doubt'. Truth is relative; how can a machine operate in an environment where truth is relative? That's why, suddenly, we enter the stage where human leadership is required.”
"We should realise that it's not the Cold War, where you had… the Berlin Wall” - he divides the table with a flat hand and points to either side of it, “free, unfree. We are dealing with a world of very blurry lines.
Governments in the UK, in France, in Germany, in the US, must recognise that this threat does exist. In Germany, for three years since the beginning of the refugee crisis, Putin's propaganda has been beefing up the AfD (German right wing). Giving them support. They [the German government] knew about it. But because German businesses are in Russia, the Social Democrats are part of the coalition – Merkel did nothing. The end result? Ninety-four neo-Nazis in the parliament.”
What’s most frustrating to Kasparov is that “it's not that we lack the means to respond. It's about political will. We have to recognise that there will always be people who will try to use [cybercrime] against the free world, so we always have to be one step ahead. It's not about slowing down the competition, we're actually speeding it up, just to make sure we are at the cutting edge. I believe in the winning spirit, the pioneer spirit, of the people of the free world. There will always be ideas coming in - we have to make sure that the influx of ideas doesn't stop.”
Kasparov describes China as “a more long-term threat. It doesn't meddle in elections. It is stealing, not for immediate use, but maybe later on. But you have so many other potential players, weaker players that could say 'Putin did it, why don't we do that?'”
“The key issue for me,” says Kasparov, is “the correlation between privacy and security… I was born and raised in the Soviet Union. I lived in Putin's Russia. I am very suspicious of government authority, and I could see many times how exceeding authority was turning into abuse of power.”
In major cybercrimes such as the recent Equifax hack, which investigators have said bears the hallmarks of a state actor, he says that “nation states, like Russia or China, look for easy prey. But that's as a result of negligence, so we have to hold our administrations, in America or Europe, responsible for not paying enough attention to defend this data.
“Millions of personal files have been stolen in America? How come? This is supposed to be the leading country, it has Google and Microsoft and Facebook, and it doesn't know how to defend itself. It's because governments are way behind. The free world has been very slow in recognising that this is the new front line. This is where you have to fight, this is where you have to invest.”
You Might Also Read:
Five Things AI Can Do Better Than Humans:
A Major Development in Deep-Learning:
AI And Robotics Can Fight Cyber Crime: