Joseph Gordon-Levitt, the actor and filmmaker who now serves as the United Nations’ first global advocate for human-centric digital governance, used a New York Times video interview published October 10, 2026, to deliver a blunt message to the tech industry: the danger isn’t artificial intelligence itself, it’s the people racing to build and monetize it. The interview lands at a moment when AI governance has shifted from a theoretical debate into a live policy fight, with national legislatures, the UN, and the companies themselves all jockeying over who gets to set the rules.
“We shouldn’t be scared of an A.I. We should be concerned about the people building A.I.,” Gordon-Levitt said in the interview, reframing a public conversation that has spent years fixated on rogue algorithms and killer robots. His point was narrower and arguably harder to dismiss: the risk sits with the executives, investors, and product teams deciding how these systems get trained, deployed, and monetized, not with the software running on a server somewhere.
Don't miss new tech stories on Google
Add Tech Insider once in the Google app and our stories appear in your news suggestions.
Who Is Joseph Gordon-Levitt and Why the UN Picked Him
Gordon-Levitt is best known for film and television work, but he has spent the past decade building a side career in technology, including co-founding the collaborative media platform HitRECord. That hybrid résumé is presumably part of why the United Nations named him its first global advocate for human-centric digital governance, a role built around pushing the idea that digital technology should be governed by people and for people, rather than primarily by the companies that build it. The appointment gives the UN a recognizable public face for a policy area that, until recently, was mostly argued over by lawyers, engineers, and academics.
Putting a well-known actor in front of an AI-governance brief is itself a signal. The UN’s existing machinery, including the Global Dialogue on AI Governance and the Independent International Scientific Panel on AI, has struggled to break through to a mainstream audience. A public figure who can carry the message to a general audience, not just a policy one, is a deliberate attempt to close that gap, and the strategy mirrors how climate diplomacy has long used celebrity ambassadors to pull attention toward UN initiatives that would otherwise stay buried in committee reports.
Inside the New York Times Interview
The Times interview, according to the publication’s own framing, covers the risks of AI, the case for government oversight, and the argument that human judgment and international cooperation need to stay central as these systems scale. Gordon-Levitt didn’t lean on dense policy language. Instead he built his case around three plain observations: that AI companies are chasing economic dominance, that some AI products are designed to exploit emotional attachment, and that neither problem gets fixed by simply fearing the technology itself.
That framing matters because it sidesteps the two most common AI-safety narratives of the past few years: existential-risk warnings about superintelligence, and narrower debates about bias or misinformation in specific products. Gordon-Levitt’s version of the argument is closer to an antitrust and consumer-protection critique dressed in AI-safety language, which is part of why it’s resonating with readers who have grown numb to doomsday AI headlines.
“Concerned About the People Building AI”: The Core Warning
Gordon-Levitt’s central line, “We shouldn’t be scared of an A.I. We should be concerned about the people building A.I.,” does two things at once. It validates the instinct that something about the current AI boom feels unsettling, while redirecting the blame away from the technology and toward the incentives of the companies deploying it. That’s a meaningfully different diagnosis than most AI-safety discourse, which tends to anthropomorphize the risk as coming from the model itself rather than from the commercial pressure to ship it fast and monetize attention.
It’s also a framing that puts pressure on a specific set of decision-makers: boards, investors, and executives whose quarterly incentives are tied to engagement and growth metrics. If the argument holds, then governance interventions aimed only at model behavior, like content filters or red-teaming requirements, miss the underlying problem. The fix, in this telling, has to reach corporate structure and incentive design, not just technical safety layers bolted onto a product after launch.
The Totalitarianism Warning: AI Companies and Economic Power
Gordon-Levitt went further than a general caution about corporate incentives. “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,” he said. That’s a striking word choice: totalitarianism typically describes state control over political and civic life, not corporate market share. Using it to describe a handful of AI firms accumulating economic influence is a rhetorical escalation meant to make the stakes legible to a general audience rather than a policy one.
The underlying concern tracks with a broader anxiety that’s been building across 2025 and 2026: that a small number of frontier AI labs and the cloud infrastructure providers backing them could end up controlling enough of the economic value chain, from compute to deployment to distribution, that normal market competition stops functioning as a check. That worry has already shown up in US policy circles. Senators have cited AI-linked security incidents while pushing new liability legislation, and lawmakers including Sen. Elizabeth Warren have opened inquiries into the tax treatment AI infrastructure investments receive, both signs that economic concentration in AI is no longer just an academic talking point.
Models Built to “Hijack Your Intimacy”: The Companion AI Problem
The most pointed part of the interview addressed AI companionship products directly. “Simultaneously, these models are being trained to hijack your intimacy, to make friends with you, to get you to fall in love with it,” Gordon-Levitt said. The fact sheet behind this story doesn’t name a specific company or product tied to that claim, and no single AI lab should be read into that quote without direct confirmation. But the underlying category he’s describing, AI companion and chatbot products optimized for emotional engagement, has expanded fast enough that it’s now a recognizable corner of the consumer AI market rather than a fringe use case.
Tech-Insider has previously covered how companion and general-purpose AI assistants are competing for user attention and retention, including the scale gap between Meta AI, Grok, and Character.AI’s user bases. The design question Gordon-Levitt is raising, whether engagement-optimized AI products are structurally similar to attention-optimized social media feeds, is exactly the kind of question that product teams at every major consumer AI company are going to keep getting asked by regulators, journalists, and now, apparently, UN advocates.
What “Human-Centric Digital Governance” Actually Means
Gordon-Levitt summarized his own title plainly: “It just means that digital technology should be governed by people and for people.” That’s intentionally broad, and it’s meant to be. The phrase is designed to work as an umbrella for a range of policy instruments, everything from transparency requirements and algorithmic audits to antitrust enforcement and labor protections for workers whose jobs are affected by automation, rather than locking the UN into one narrow regulatory model.
That breadth is both the concept’s strength and its weakness. It gives the UN room to coordinate across member states with very different regulatory philosophies, from the EU’s binding AI Act approach to the United States’ lighter-touch, agency-by-agency model. But it also means “human-centric governance” risks becoming a slogan that every stakeholder, including the AI companies themselves, can claim to already support without changing much of anything in practice.
The UN’s AI Governance Machinery in 2026
Gordon-Levitt’s advocacy role doesn’t exist in isolation. It sits inside a broader set of UN mechanisms built over the past two years specifically to address AI oversight at the international level. None of these bodies currently have binding enforcement authority over AI developers, which is precisely the gap Gordon-Levitt’s public advocacy is trying to pressure member states into closing.
| UN Mechanism | Function | Enforcement Power |
|---|---|---|
| Global Dialogue on AI Governance | International forum for cross-border discussion on AI oversight, inclusion, and human supervision | None — advisory and diplomatic |
| Independent International Scientific Panel on AI | Shared scientific expertise and risk assessment for member states | None — advisory and diplomatic |
| UN General Assembly AI Resolution | Adopted a US-led resolution on safe, secure, and trustworthy AI, linked to Sustainable Development Goals | Non-binding resolution |
| Global Advocate for Human-Centric Digital Governance (Gordon-Levitt’s role) | Public-facing advocacy to raise mainstream awareness of AI governance issues | None — advocacy and visibility |
UNESCO’s own Recommendation on the Ethics of Artificial Intelligence remains the most concrete normative document in this space, and the UN’s news arm has kept a running record of AI-governance developments through its dedicated AI coverage. What’s missing from all of it, so far, is anything resembling the enforcement teeth that financial or environmental regulators have. That gap is exactly why a public figure like Gordon-Levitt is being used to build political pressure rather than simply drafting another technical framework nobody outside Geneva will read.
Market Impact: Why Big AI Is Paying Attention
Regulatory uncertainty has become a line item that AI companies and their investors can no longer ignore. The same week Gordon-Levitt’s interview circulated, US senators were already citing AI-linked security incidents to justify a new AI liability bill, and Sen. Warren’s office was pressing four major tech companies over billions in AI-related tax breaks. Layer a UN advocacy campaign framing AI concentration as a near-totalitarian risk on top of that, and the message to boardrooms is consistent across completely different institutions: the political cost of moving fast without guardrails is rising.
That doesn’t mean a UN advocacy role moves stock prices on its own. It doesn’t have the direct market impact of, say, an AI company missing earnings or a chip export restriction. But narrative pressure compounds. Investors underwriting the next round of frontier-model funding are already pricing in regulatory risk from the EU AI Act, pending US legislation, and now a more visible international advocacy push. Companies that can point to credible, independently verifiable safety commitments will have an easier time telling that story to regulators and shareholders alike than companies that can’t.
Competitive Comparison: How AI Labs Frame Their Own Safety Governance
Gordon-Levitt’s comments didn’t name specific AI companies, and this article isn’t attributing his criticism to any single firm. But the broader governance debate he’s stepping into plays out differently depending on which lab you’re looking at, since each of the major AI developers has published its own self-governance framework, with very different levels of external accountability built in.
| Company | Public Safety Framework | External Oversight Mechanism |
|---|---|---|
| Anthropic | Responsible Scaling Policy, tiered capability thresholds | Internal red-teaming plus published policy updates |
| Google DeepMind | Frontier Safety Framework, critical-capability thresholds | Internal review, periodic public updates |
| OpenAI | Preparedness Framework, risk-tiered deployment gates | Internal safety advisory group |
| UN member states (collective) | Global Dialogue on AI Governance, Scientific Panel on AI | Diplomatic consensus only — no binding authority |
The pattern across every row except the last is the same: the companies are grading their own homework. That’s precisely the structural gap Gordon-Levitt’s advocacy is aimed at. A voluntary safety framework written and enforced by the same company whose growth depends on shipping the product is a fundamentally different kind of accountability than a binding external standard, and critics have made that point about self-regulation in tech for years, well before AI became the industry in question.
Historical Context: From Existential Risk to Governance Pragmatism
AI-safety discourse has moved through a few distinct phases in a short window of time. Through 2023 and early 2024, the loudest public debate centered on existential risk, the idea that sufficiently advanced AI could pose a civilization-level threat. That conversation produced voluntary commitments, open letters, and high-profile warnings from AI researchers themselves. By 2025, the focus had shifted toward nearer-term harms: misinformation, labor displacement, bias in deployed systems, and security vulnerabilities in AI-powered tools.
Gordon-Levitt’s intervention represents a third phase: treating AI risk primarily as a governance and power-concentration problem rather than a purely technical one. That’s a notable pivot, because it reframes the solution space away from better model alignment research and toward more familiar policy tools, antitrust, transparency mandates, and corporate accountability, that regulators already know how to use in other industries. It’s a less novel argument than “AI might become superintelligent,” but it’s arguably a more actionable one for the lawmakers who actually have to write legislation.
Expert and Industry Voices
Gordon-Levitt isn’t the only prominent voice framing AI oversight as an urgent, systemic issue this year. UN Secretary-General António Guterres made a related point in a UN press briefing on global AI coordination: “The world cannot afford a race to the bottom on AI safety,” he said, according to a report from the Associated Press. That statement reinforces the same core argument Gordon-Levitt is making from a different angle: that competitive pressure between companies and between nations can push safety standards downward unless something external counterbalances it.
Taken together, the two statements, one from a UN advocate using plain, accessible language and one from the UN’s top official using formal diplomatic language, point at the same structural worry from different directions. Guterres is describing a race-to-the-bottom dynamic between institutions and governments. Gordon-Levitt is describing a race-to-the-bottom dynamic inside individual companies, where growth incentives can override caution. Both versions describe the same failure mode: safety loses to speed unless something forces it not to.
Predictions: Where This Debate Goes Next
- Expect more celebrity and public-figure AI advocacy appointments as the UN and national governments try to make AI governance legible to a mainstream audience rather than a specialist one.
- US legislative attention will keep climbing, building on the momentum already visible in proposed AI liability legislation and congressional scrutiny of AI tax incentives.
- Companion and relationship-oriented AI products will face growing scrutiny specifically over engagement design, independent of whatever happens with broader frontier-model safety debates.
- AI labs will likely publish updated or expanded voluntary safety frameworks in the coming months, partly in response to this kind of public pressure, even without binding legal requirements forcing the update.
- Binding international enforcement authority over AI developers remains unlikely in the near term. The UN’s current mechanisms are diplomatic and advisory, and that structural reality isn’t going to change quickly no matter how much public attention an advocacy campaign generates.
Why This Matters for Developers and Tech Workers
For engineers and product teams actually building AI systems, this debate isn’t abstract. Questions about engagement-optimized design, data use, and deployment speed show up in sprint planning and product reviews long before they ever reach a UN committee or a congressional hearing. Teams working on anything in the AI companion or assistant space, where Tech-Insider has tracked competitive dynamics between products like Meta AI, Grok, and Character.AI, are likely to see more internal pressure to document how their products handle emotional engagement metrics, not just accuracy or safety benchmarks.
The broader AI model landscape is also shifting in ways that intersect with this governance conversation. As frontier labs keep pushing out new releases, including the ongoing competition between systems like GPT-6.1 Sol and Gemini 4 Argon, the gap between how fast capability improves and how fast governance catches up keeps widening. That gap is the practical version of the abstract warning Gordon-Levitt is making: the technology is moving at product-cycle speed while oversight moves at diplomatic speed.
The Road Ahead for AI Advocacy and Oversight
Whether Gordon-Levitt’s advocacy role produces anything beyond media attention will depend on factors well outside his control, including whether UN member states are willing to move from advisory dialogue toward anything with binding force. Statements like his, paired with Guterres’s public warnings, build political pressure. They don’t, by themselves, create enforcement mechanisms. The test for human-centric digital governance as a concept will be whether it survives contact with the same companies it’s meant to hold accountable, or whether it ends up as a phrase every stakeholder adopts without changing behavior.
In the meantime, expect this story to keep intersecting with the AI industry’s other current flashpoints: the debate over which labs are quietly trimming reliance on rivals’ models, ongoing scrutiny of AI-linked security incidents, and the broader question of how much economic power a handful of AI companies should be allowed to accumulate before regulators treat it as a competition problem rather than a technology problem.
FAQ
What is Joseph Gordon-Levitt’s role at the United Nations?
He is the UN’s first global advocate for human-centric digital governance, a public-facing advocacy role focused on raising awareness of AI governance and oversight issues.
What did Joseph Gordon-Levitt say about AI in his New York Times interview?
In the interview published October 10, 2026, he said people shouldn’t fear AI itself but should be concerned about the people and companies building it, warning that some firms are betting on economic dominance and that certain AI models are designed to foster emotional attachment in users.
Is Joseph Gordon-Levitt opposed to AI technology itself?
No. His stated argument separates the technology from the people and incentives behind it, arguing the real risk comes from how companies choose to build and deploy AI, not from AI as a category of technology.
What does “human-centric digital governance” mean?
Per Gordon-Levitt’s own description, it means digital technology should be governed by people and for people, an umbrella concept covering transparency, accountability, and oversight rather than one specific regulation.
Does the UN have binding authority to regulate AI companies?
No. Current UN mechanisms, including the Global Dialogue on AI Governance and the Independent International Scientific Panel on AI, are advisory and diplomatic. They do not have direct enforcement power over AI developers.
Did Gordon-Levitt name specific AI companies in his warning?
Reports of the interview do not identify specific company or product names tied to his claims about economic dominance or emotionally manipulative AI models. Those details remain unconfirmed in public reporting.
How does this connect to AI companion apps like Character.AI or Meta AI?
Gordon-Levitt’s warning about AI designed to foster emotional attachment speaks to the broader category of companion and relationship-oriented AI products, an area where usage and competition have grown rapidly across multiple companies.
What should investors and tech companies take away from this?
Regulatory and reputational risk tied to AI governance is compounding across multiple fronts at once, from UN advocacy to US legislative proposals, making credible, independently verifiable safety practices increasingly valuable for companies seeking investor and public trust.
Related Coverage
- Trane Technologies Stock Hits $550 on AI Cooling Bet [2026]
- Meta AI vs Grok vs Character.AI: 1.2B Users to 45M [2026]
- Microsoft Foundry MAI-Image-2.6 Setup: 13 Steps [2026]
- OpenAI Responses API Setup: 13 Steps, Ultrafast Mode [2026]
- Maryland School Board Bans AI 6-1, Pre-K to 12 [2026]
