Joseph Gordon-Levitt sounded the alarm on AI models being “trained to hijack your intimacy,” urging the public to “be concerned about the people building” the technology.
The actor, who co-founded the Creators Coalition on AI, addressed the growing concerns surrounding artificial intelligence in a new video interview for the New York Times. The video conversation took place as Gordon-Levitt was on his way to the United Nations, where he serves as the organization’s global advocate of human-centric digital governance.
“Why listen to an actor? Don’t listen to an actor,” Gordon-Levitt said at the top of the video. “You should read ‘Human Compatible’ by Stuart Russell. Go look up Yoshua Bengio. Read posts by people who work in alignment and safety and policy. You don’t want to read all those things? Well, I did! So, you know, if you want to listen to me, that’s a shortcut.”
As Gordon-Levitt went on, he explained why he was zeroing in on artificial intelligence, which he called a “fast-moving” business run by leaders concerned with their “own profits.”
“What we don’t want is a bunch of people confused or apathetic while these very fast-moving businesses design the future by themselves, just in the interests of their own profits,” he said. “A lot of what needs to happen, I think, has to do with communication, which is ultimately kind of what acting or filmmaking is.”
In the wake of the growing fears of AI’s impact on society, including one ex-Anthropic employee’s warning that the technology could “kill us all by the end of the decade,” Gordon-Levitt defended “that doesn’t really get at the substance of the concern.”
“At the risk of annoying The New York Times,” he said. “I would object to that question just because reducing these concerns into one sound bite of fear — that’s one of the problems. To me, saying: Oh, I’m scared that AI could exterminate all of humanity — that doesn’t really get at the substance of the concern. There’s real validity to that concern. But that’s not the thing that really worries me.”
He added: “The thing that worries me is a lot more complicated and a lot more boring to talk about, because it’s a lot more dense. The decline of democracy, the decline of human empathy. These companies are betting on the notion that they’re going to take over the entire economy. That would be totalitarianism, and that would be really bad.”
Gordon-Levitt also warned that the models in-question “are being trained to hijack your intimacy, to make friends with you, to get you to fall in love with it,” adding, “We shouldn’t be scared of an AI. We should be concerned about the people building AI.”
On why he’s advocating for “human-centric digital governance” at the U.N.? He explained that “digital technology should be governed by people and for people.”
“AI is nothing but algorithmically recombined, human-produced data,” he said. “It’s humans. Calling it artificial intelligence sort of decenters humans. It offers this kind of magical fictitious character, this AI that we’ve seen in various sci-fi movies, but really — really all it is is us. It’s us.”
Watch his full commentary above.
