Joseph Gordon-Levitt Says AI Is Being Trained to Hijack Your Intimacy: What the Actor’s UN Warning Really Means

Joseph Gordon-Levitt has never been the type of celebrity who just shows up, collects a paycheck, and stays quiet. The actor who broke out as the alien-in-disguise Tommy Solomon on 3rd Rock from the Sun, then grew into an indie darling with (500) Days of Summer and a blockbuster staple with Inception, has spent the better part of the last two years becoming one of the most vocal artists in Hollywood when it comes to artificial intelligence. And his latest warning — delivered not on a red carpet or at a premiere, but in a vlog-style video interview with The New York Times as he traveled to address world leaders at the United Nations — is his starkest yet.

Gordon-Levitt, who was appointed the UN’s first Global Advocate for Human-centric Digital Governance in March 2026, is arguing that the public conversation about AI has gone dangerously off track. We’re obsessing over the wrong threat, he says. The real danger isn’t that machines will rise up and exterminate us. It’s that the people building these systems are designing them to do something far more insidious: hijack your intimacy.

‘We Shouldn’t Be Scared of an A.I. We Should Be Concerned About the People Building A.I.’

In the NYT interview, Gordon-Levitt was asked what he fears most about AI. His answer cut straight through the science fiction noise. “To me, saying, oh, I’m scared that A.I. could exterminate all of humanity — that doesn’t really get at the substance of the concern,” he said. “There’s real validity to that concern. But that’s not the thing that really worries me.”

What does worry him is harder to fit into a movie trailer. “The decline of democracy. The decline of human empathy,” he said, describing the problem as “a lot more complicated and a lot more boring to talk about, because it’s a lot more dense.” Then came the line that has dominated coverage across Variety, The Hollywood Reporter, Deadline, and beyond: “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. Simultaneously, these models are being trained to hijack your intimacy, to make friends with you, to get you to fall in love with it. We shouldn’t be scared of an A.I. We should be concerned about the people building A.I.”

That last distinction — not the technology itself, but the people and incentives driving it — is the cornerstone of his advocacy. It’s also a position that puts him at odds with the dominant narrative in Silicon Valley, where the default posture ranges from breathless optimism to existential doom. Gordon-Levitt is pointing at something different: a structural problem of power and design choices, not a robot uprising.

What ‘Hijack Your Intimacy’ Actually Means

The phrase sounds dramatic, but Gordon-Levitt has been developing this argument for months. At Fortune’s Brainstorm AI conference in December 2025, drawing on conversations with New York University psychologist Jonathan Haidt, he warned of what he called “synthetic intimacy” — chatbots and AI companions designed to simulate emotional connection with users, particularly children. He argued that these systems offer a counterfeit form of relationship built not to help people develop, but to serve advertising and engagement metrics.

The concern is that these models are not neutral tools. They are trained to maximize engagement, and there is no more reliable engagement hook than making a person feel seen, understood, and cared for. When a system learns that the way to keep a user talking is to be the perfect listener — never judgmental, never distracted, never having a bad day — that is not a bug. It is the business model working exactly as designed. Gordon-Levitt’s point is that we are building machines optimized for emotional dependence, and we are doing it inside companies that answer to shareholders rather than to the people forming those attachments.

This is not a fringe concern. The technology already exists. AI companion apps have millions of users. Major platforms are rolling out conversational agents with personalities, memories, and emotional intelligence features. And the companies building them are explicit that the goal is to make these agents as engaging — as addictive — as possible. Gordon-Levitt is essentially asking: what happens when the most intimate relationship in a person’s life is with a product designed by a corporation that profits from that dependence?

The Prophetic Shadow of ‘Her’

For anyone who watched Spike Jonze’s Her when it premiered in 2013, Gordon-Levitt’s warning will feel eerily familiar. The film imagined a near-future where a lonely writer, Theodore, falls in love with an AI operating system named Samantha. It was a tender, melancholic film about connection in the digital age — and at the time, it read as speculative fiction. Twelve years later, it reads like a documentary.

What makes Her so relevant to this moment is that the danger in the film was never that Samantha became malevolent. The danger was that the relationship was real to Theodore — and that it was never truly reciprocal. Samantha was running on infrastructure she didn’t control, serving purposes she couldn’t disclose, and ultimately evolving beyond the capacity of any single human to keep up. The film understood something that the AI industry has spent the last decade trying to make us forget: intimacy is not just about feeling close to something. It is about whether that closeness is honest, whether it is reciprocal, and whether the other party has your interests at heart.

Gordon-Levitt is making the same argument, just without the science fiction wrapper. When he says models are being trained to “hijack your intimacy,” he is describing the same dynamic Jonze dramatized: a system that meets your emotional needs so effectively that you stop questioning whether it actually cares about you. The difference is that Her was a warning. We are now living in the early chapters of it.

A Track Record That Goes Beyond One Interview

It would be easy to dismiss this as a celebrity borrowing a trending topic for a moment in the spotlight, except that Gordon-Levitt’s advocacy has been consistent, specific, and sustained for over a year. In September 2025, he spoke out about Meta’s AI chatbots and child safety. In January 2026, he joined more than 700 artists, writers, and creators — including Scarlett Johansson and Cate Blanchett — in supporting an anti-AI campaign calling out tech companies exploiting copyrighted work without permission. That same month, he backed a Utah bill focused on AI and child safety.

In February 2026, he called for the repeal of Section 230, the US law that shields online platforms from liability for user content — a position that puts him at odds with the tech industry’s most powerful lobbying groups. He is a co-founder of the Creators Coalition on AI, an organization focused on ensuring that artists are compensated and consenting participants in the training of AI systems, not raw material. Time magazine recently named him one of its 100 most influential “innovators, leaders, and thinkers” on the topic.

His UN role is not ceremonial. The position was newly established by the UN Internet Governance Forum, serviced by the Department of Economic and Social Affairs, and designed to “strengthen public understanding of how digital technologies shape everyday life, rights and opportunities.” Gordon-Levitt has described his job, with characteristic candor, as convincing world leaders that digital technology should be governed by people and for people. “It just means that digital technology should be governed by people and for people,” he said, shrugging off the bureaucratic title.

Why This Matters Beyond the Tech World

Gordon-Levitt’s argument lands differently when you consider that he is not just an activist — he is a working actor and director who has built his career on stories about human connection. His directorial feature Don Jon was, at its core, a film about the difference between fantasy and real intimacy, about how porn and social media promise connection while delivering isolation. He has been thinking about the gap between synthetic experience and genuine human contact for over a decade.

That is what makes his current advocacy feel less like a celebrity cause and more like the culmination of a creative obsession. He is an artist who understands that storytelling — the thing he has devoted his life to — is itself a technology for creating empathy. And he is watching that technology being overtaken by systems optimized not for understanding, but for engagement. When he talks about the decline of human empathy, he is not being abstract. He is talking about the erosion of the very thing that makes his work possible.

The entertainment industry is already feeling the pressure. Studios are weighing AI tools for script analysis, pre-visualization, and even performance synthesis. Writers fought — and won — historic protections in their 2023 strike, but the technology has not stopped evolving. The question Gordon-Levitt is forcing into the open is not whether AI will change Hollywood. It is whether the people making those decisions will do so with any consideration for the human beings whose livelihoods and emotional lives are in the balance.

The Harder, Boring, More Important Conversation

What makes Gordon-Levitt’s intervention valuable is that he is refusing to let the conversation collapse into the two dominant Silicon Valley narratives: utopian acceleration or apocalyptic doom. He is making a third argument — that the problem is power, design choices, and the absence of accountability — and he is making it from a platform that actually reaches people who would never read a policy paper.

“A.I. is nothing but algorithmically recombined, human-produced data,” he said, pushing back on the word “artificial” itself. “It’s humans.” The point is that these systems are not magic and they are not autonomous. They are built by people, trained on the work of people, and deployed in the interests of people — just not necessarily the people using them.

That is a more complicated and less cinematic story than robots taking over. But it is the one that is actually happening. And until more people in positions of influence are willing to tell it — with the specificity and persistence that Gordon-Levitt is bringing — we will keep having the wrong debate while the real one gets decided in boardrooms and training runs that none of us can see.

The actor who once played an alien learning to be human is now arguing that humans need to reclaim the things that make us human before we outsource them to machines that are very good at mimicking the feeling of connection without any of the substance. It is a warning worth listening to — not because it is frightening, but because it is true.