You had to feel for moderator Jason Tanz and academic Francesca Rossi during a panel this week at the Paley Center for Media on how AI is changing the world.
Wired’s Tanz and IBM Research’s Rossi were sharing a stage with former world chess champion and AI scholar Garry Kasparov, and like most of his opponents over the chessboard in the past they were discovering that his tends to be the biggest brain in the room, whichever room in the world that might be.
Emanuel Lasker, another former world chess champion, once said: "On the chessboard lies and hypocrisy do not survive too long," and Kasparov has taken that edict into his post-chess career in politics, human rights advocacy - and AI. His latest book is entitled Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins.
The idealistic Rossi envisaged a semi-utopian future where every AI system is checked for bias by reputable thirty-party verifiers, and cited her work with the 'Partnership on AI (to benefit people and society)' as a vehicle for multidisciplinary partners who produce and are impacted by AI to come together for the greater good.
Just as in his chess career, Kasparov was an open book when it came to displaying his thoughts on a statement during the discussion, with exaggerated head shaking, incredulous expressions, and apparent disdain for the lesser mortals around him.
It’s not deliberate, it’s just that Kasparov is super-intelligent, extremely well informed, and incredibly passionate about his beliefs – whatever the topic.
He responded to Rossi’s optimism by noting that AI is in essence a technology, one that can be used for good or evil - and is always easier to use for evil.
"To destroy is easier than to build," he added. "I cannot accept that AI will amplify the best of humanity. It will emphasize the worst of humanity, as every technology has before. It’s great for building financial and social networks, but it’s also good for building a sophisticated terrorist network."
He believes we are entering a new world where very powerful – and cheap - technology can be totally destructive and used to poison people’s minds. "The idea that by making these technological improvements we enhance humanity and the human race doesn’t stand the test of history," he added.
Kasparov’s view is that attempts to create regulation and certification will work in the free Western world but will be ignored in places like Russia, China, Iran, and Turkey.
"Many places will look at AI as a great opportunity to spread chaos and advance their clandestine agenda by using AI and the brains behind AI in the free world," said Kasparov. "The U.S., Italy, and Japan will abide by the values of the free world - others will not."
Even in the West, Kasparov pointed to double standards. "Look at Google and Apple," he said. "In the U.S. they play by the rules, but to protect their businesses in Russia and China they are willing to bend the rules. They will defend individual rights and data in this country, and they’re willingly conceding this information to the KGB or Chinese intelligence, endangering millions of people."
As long as multinational corporations practice double standards in the free world and non-free world, Kasparov believes all talk about AI and partnership is pie in the sky.
"It’s good for people who were born and raised in the free world, but what about those in Russia and China who want to benefit from technological progress but know if they share any info it could end up in the hands of the KGBs of this world they might pay an ultimate price," he said.
Rossi cited the likes of Human Rights Watch and UNICEF as protectors who can make sure things are done in the right way everywhere. Kasparov brutally noted: "They all do a great job, but the KGB doesn’t care about them. Oligarchs in Russia and China don’t care about it. If we want to see progress, Google, IBM, and Apple need to follow regulations worldwide."
"This is depressing," said discussion moderator Tanz, as he attempted to close the session on a more positive note and move to a discussion about jobs in an AI future. That didn’t work either.
"Any great innovation such as AI will destroy jobs and create jobs," said Kasparov. "My concern is not that we are moving too fast, but we are moving too slow. These jobs are going, and if we try and protract the agony it will slow down the process and we will not reach the end of the curve to create the industries and provide economic growth and provide enough financial support for those who are left behind."
As he pointed out: "This has been happening with manufacturing jobs for decades, if not centuries. The only problem is that AI is now going after people with college degrees and Twitter accounts."
Rossi was still trying to be upbeat. She shared some findings from a conference she attended on ethics and society: "90% of jobs that were active in the 1900s do not exist now. But still unemployment is 4%, so many, many new jobs were invented."
The changes are much faster with AI and all agreed that, as technology replaces work, we are pushed harder to discover what it is that only humans can do.
Kasparov said: "In 2016, in the U.S. job market, only 4% of jobs required creativity. Basically, 96% of jobs could be automated. We don’t recognize how many jobs today just don’t require true human qualities."
Kasparov is famous for being the first chess world champion to lose to a computer, when he went down against IBM’s Deep Blue – the forerunner of Watson – in a controversial match in 1997, although he was quick to point out he won the first match they played, in 1996.
It was seen as a seminal moment in the "war" between man and machine, and a tipping point that presaged a new era in which computers had finally overtaken humans in terms of intelligence.
"I was quite upset when I lost the match in 1997," said Kasparov. "I didn’t lose very often. I was angry with myself because I played way below my level. I made a lot of mistakes and my performance was lousy."
Former world champion Garry Kasparov feels the full power of chess played by machines.
But chess didn’t die, despite the fact you can now get a standard program on your cellphone that is stronger than Deep Blue was back when it defeated Kasparov - in fact they are 500 ELO rating points stronger than current chess world champion Magnus Carlsen.
As Kasparov notes, ultimately it is about humans competing against humans. There are many animals that run much faster than humans, but that hasn’t ended the interest in finding out who is the fastest human on earth in the 100 meters final at the Olympics.
Back then, Deep Blue could calculate 200 million positions per second but even with this incredible capability it could not solve the game, which is mathematically almost infinite in its possibilities. Rather, Kasparov said, it meant Deep Blue made fewer mistakes than its opponent and thus achieved the real objective, which was winning the game – not solving it.
He referenced the distinction between type A machine intelligence, which is about brute force and algorithms, and type B, which somewhat emulates elements of human thinking. Type B machines never took off in chess, which has ultimately been conquered by machines using type A brute force.
On the day when Elon Musk’s SpaceX rocket was propelled into space, Kasparov reflected that kids in the 1960s all wanted to be space engineers, but by the 1990s their passion had turned toward finance.
"It was the wrong track because we lost our passion for space," said Kasparov. "I sincerely hope AI will help transfer jobs from the financial industry to space."
Kasparov says humans must swallow their egos and pride and realize there are many things machines do better. And this doesn’t apply just to chess.
Tanz cited the example of self-driving cars: it’s not getting the cars to drive, it’s alerting the humans to prefer to use them. Kasparov noted that, in 25 years, our kids and grandchildren will discuss how stupid their parents were to drive cars.
"Cars are one of the great killers of our time," he said. "We know that if anything goes wrong with a self-driving car it’s on the front page of a newspaper, but at the same time 1,000 people could be killed in car accidents and that’s taken for granted."
AI is a relatively cheap and universally available technology, especially compared to something like nuclear power. Any optimism in the room at the end of a fascinating but sometimes gloomy discussion seemed to solidify around a feeling that AI could be the catalyst for a new generation of creativity, where fresh companies founded in people’s garages rise up to compete with the likes of Apple and Google.
In January, the 2018 Bloomberg Innovation Index showed that, for the first time in the six years the survey has been published, the U.S. was ranked outside the top 10 – coming in 11th out of 50 economies - so this new generation of AI-inspired creativity and innovation can't come soon enough.
I left the Paley Center reflecting that it would be great to have Kasparov's big brain pointed in the direction of this nascent opportunity for innovation if the U.S. economy is to truly benefit from the game-changing potential of AI.