Skip to main content

Is Artificial Intelligence Being Oversold?

A debate in New York City tries to settle the question

This article was published in Scientific American’s former blog network and reflects the views of the author, not necessarily those of Scientific American


After taking a 2-0 lead in its five-game match with Lee Sedol on Thursday, Google-DeepMind’s AlphaGo artificial intelligence program seems likely to claim victory within the next few days. This will no doubt resurface the many questions people have about AI’s future and whether humans are inching towards Matrix-like enslavement. Fortunately, last night’s "Don’t Trust the Promise of Artificial Intelligence" Intelligence Squared U.S. debate in New York City addressed a lot of these questions and concerns.

Arguments both for and against the motion were complex and nuanced but basically it was a debate over whether AI is being oversold at this time. “Go games are informing a Go algorithm,” computer scientist Jaron Lanier noted in his opening statements in reference to Sedol’s match against a computer. Yet Sedol isn’t just playing against one artificial intelligence program. He’s playing against the aggregate of 30 million moves made during previous games by human players. Lanier drew parallels with real-time Skype translation. “The only way we do it is by scraping the efforts of millions of translators who don't even know what's happening to them to get the examples.”

Lanier, the only debater to refer specifically to AlphaGo, warned against what he calls "premature mystery reduction, which is when we pretend we have something working that we really don't.” Most AI software today is programmed using massive datasets representing some sort of human activity, he added. “We want to pretend that there's this AI behind the curtain that's freestanding. But, actually, there's millions of people there, too.”


On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


Lanier teamed with Internet entrepreneur Andrew Keen on the side of not trusting the promise of AI, emphasizing that the debate was about “the promise” as opposed to AI itself. They cautioned against machines taking away certain freedoms by eliminating jobs and gathering increasing amounts of personal information. On the other side, James Hughes, executive director of the Institute for Ethics and Emerging Technologies, and Martine Rothblatt, CEO of biotech firm United Therapeutics, were more enthusiastic about AI’s potential, taking the position that it could free us from human frailty and “wage slave jobs”—as Keen put it—that computers can do better. All of the debaters—each of whom has a long list of credentials that you can see by following the links above—agreed that popular science fiction movies and TV do the field of AI a disservice when they portray dangerous cyborgs run amuck.

That’s of course an oversimplification of last night’s hour-long exchange of ideas, which you can see in its entirety here. For some additional perspective, The New York Times likewise posted a debate on AI’s future.