FYI.

This story is over 5 years old.

Tech

How to Turn NPR Fans Into Artificial Intelligence Cynics

The live debate series Intelligence Squared US takes on the proposition: “Don’t trust the promises of artificial intelligence.”
Intelligence Squared's host, John Donvan. Image: Intelligence Squared US

At a debate Wednesday night co-hosted by Intelligence Squared US and the 92 Street Y in New York City, a largely white, middle-aged audience was easily convinced that the supposed bright future of artificial intelligence is perhaps not all that it's cracked up to be.

The nonprofit debate series, syndicated as a podcast on NPR, works something like this: Before the debate begins, the audience votes in favor of the motion, against the motion, or as undecided. Last night's motion, "Don't Trust The Promises of Artificial Intelligence," was affirmed by 30 percent of the audience and negated by 41 percent at the start of the debate. A whopping 29 percent were undecided, indicating they knew little about the topic beforehand. Votes weren't just cast by the live audience, but also by viewers online.

Advertisement

The elderly white woman next to me, who would be asleep by the end of the debate, chose to abstain. "I haven't decided, I don't know much about it," she told me. As someone who has written before about the promises of future technology, I thought I was educated enough to vote against the motion, in favor of the future of AI.

By the end of the debate, I would change my vote, and so would much of the audience. At the poll taken at the end, 59 percent voted in favor of the motion, 30 percent voted against the motion, and another 11 percent didn't learn enough to feel comfortable casting a vote. Those skeptical of AI were declared the winners. The online results were slightly different. In that poll, 40 percent (57 votes) were in favor of the motion, whereas 60 percent were against it (87 votes).

Like more than half the people who voted live, I was no longer convinced that AI was going to save the world in any meaningful way.

Debating for the motion (in other words against the promise of AI) were Andrew Keen and Jaron Lanier. Keen's best known for his belief that technology is dismantling our culture, a view crystallized in his book The Internet Is Not The Answer. He's notoriously critical of "democratizing" information platforms, like Wikipedia and YouTube. At one point during the debate, he went so far as to say "open-source is just another ideological dream."

Keen lauded the most educated professions of our society, like doctors and lawyers, and urged the audience to consider what would happen if we tried to automate them. He rightfully emphasized that we have no idea what we're going to do when many, or most, of the world's jobs are outsourced to cheaper and faster artificially intelligent machines.

Advertisement

Keen's debate partner, Jaron Lanier, made a more subtle point. He's a widely accomplished programmer himself, and was introduced last night as the "father of virtual reality," a term he supposedly coined or brought into popular use.

He pointed out "there has to be a division between the work itself—the engineering and the science—and the narrative on the other side." He believes that the problem is that people have assigned a kind of mysticism or religion to the future of artificial intelligence, that's divorced from the reality of the scientific research.

"The only thing matters is if people can use the machines that we design," he said.

Another point he emphasized is that individuals often under appreciate just how much today's artificially intelligent machines rely on vast repositories of human data. Automatic translators for example, are only able to learn to do their job because they have access to insane amounts of documents and conversations translated by real, live bilingual humans.

Image: Louise Matsakis

On the other side of the debate, there did seem to be a kind of dogma at play, as Lanier said might be the case. Martine Rothblatt, CEO of biotech firm United Therapeutics and the founder of Sirius XM, told the audience that the three words she associates the most with AI are replication, application, and fascination.

"The promise of AI is that we will be able to replicate the human mind," she said, with awe in her voice. Her infatuation with tech isn't terribly surprising, when you factor in that Rothblatt is the founder of her own religion, the Terasem Movement, which The Washington Post says is a combination of cultural Judaism, Zen-like yoga, and a "deep belief in technology." One of the principle beliefs of the religion is that death is supposedly optional.

Rothblatt and her debate partner, James Hughes, executive director of the Institute for Ethics and Emerging Technologies, tried to equate the future of artificial intelligence with the future of human progress as a whole. Their thesis was that as far as anyone can tell, during the course of humanity's history, we've always developed new technologies that have continuously improved our lives. Why would that suddenly not be the case when it comes to artificial intelligence?

Technology has certainly made certain parts of life faster, easier, and perhaps happier, but as all of the debaters pointed out, artificial intelligence can't solve complex human problems like poverty or climate change for us.

AI certainly might have the power to help us tackle these issues, but the research just doesn't support the promise that it will "make possible abundance and leisure for us all," as Hughes said.

I'm not optimistic, but I hope that he's right. Outside, I asked a group of four young people, all who work in AI, to share their votes. Three voted against the motion, in favor of AI. Maybe artificial intelligence can only seem like the answer from the inside.