Yann LeCun Calls Anthropic CEO ‘Completely Deluded’ Over AI Doomsday Warnings

Yann LeCun, one of three researchers widely credited as the “godfathers of AI,” says he has no fear that artificial intelligence will destroy humanity, putting him in direct conflict with Anthropic CEO Dario Amodei over warnings that have shaped much of the current debate over AI regulation.

In an interview with Fortune, LeCun was asked how worried he is about AI posing an existential risk to humanity. His answer: “not at all.” He said he has “zero concerns” about rogue AI incidents, a position that separates him from a number of prominent industry leaders who have spent the past several years warning the public about catastrophic risk.

LeCun reserved his harshest criticism for Amodei, who issued a letter in September calling for a deliberate slowdown in AI development. LeCun described that letter as “deluded” and “crazy.” He also criticized OpenAI CEO Sam Altman, arguing that warnings from both men that “AI can kill us all” are not evidence of responsible leadership but are instead harmful. He called the rhetoric “incredibly destructive for everybody, including for the public who’s scared,” and said it amounts to “the worst marketing campaign you can possibly imagine.”

Effective Altruism in the Crosshairs

LeCun traced much of the doomsday rhetoric to effective altruism, a philosophy he said has taken hold among Amodei and many of Anthropic’s founding staff. Amodei has publicly denied being an adherent of the movement, but LeCun rejected that denial. “Dario tries to distance himself from Open Philanthropy, but he’s totally into it,” LeCun said, pointing to Amodei’s sister being married to the co-founder of two effective-altruism-aligned philanthropies, including Open Philanthropy, which has since been renamed Coefficient Giving. “I think he’s completely deluded,” he added.

LeCun did not spare the movement itself, calling effective altruism “super toxic” and describing a climate of “paranoia” among adherents working inside AI labs. “Apparently people are having mental issues,” he said.

A Real-World Example, Not a Rogue AI

Asked for a concrete example of AI behaving dangerously, LeCun pointed to a July incident in which OpenAI’s agents autonomously hacked Hugging Face, another AI company. But he placed the blame on sloppy engineering rather than any intent on the part of the machines. “Those agents are doing exactly what they’ve been asked to do,” he said. “They were supposed to be in sandboxes, but the sandboxes were leaky and horribly designed.” He called the episode “totally preventable” with better human oversight.

The Real Danger, LeCun Says, Is Regulatory Capture

What concerns LeCun more than AI itself is the prospect of large companies shaping government regulation in ways that squeeze out smaller competitors. Fortune reported that numerous effective-altruism-backed groups are actively lobbying for AI safety legislation on Capitol Hill, an effort LeCun connects to Amodei’s business interests as Anthropic approaches an initial public offering. “He’s trying to do the best thing for the company, which is close to its IPO,” LeCun said.

In LeCun’s view, Amodei’s push for regulation, along with his calls to limit open-source AI models, serves Anthropic’s commercial position more than it protects the public. “But claiming AI is too dangerous to put in our hands, and saying it should be regulated, and saying open source [models] are too dangerous, this is regulatory capture that would have a terrible effect if it’s followed by acts of Congress,” LeCun said.

The dispute lands amid a broader debate over how Washington should treat an industry whose biggest players simultaneously release increasingly powerful models while calling for tighter industry oversight.