AI Leaders Push Self-Regulation as Existential Fears Mount

Sam Altman stood before hundreds of business leaders at a San Francisco conference on Tuesday and made a simple plea: trust us. The OpenAI chief executive acknowledged that fears about artificial intelligence are justified—the technology has advanced so rapidly that the worst-case scenarios no longer require much imagination. But he insisted that his company and others like it have both the moral compass and the competitive incentive to regulate themselves, to keep safety ahead of capability, and to slow down or stop if they cannot.

Altman's remarks came in the wake of a viral post by a researcher who had quit Anthropic, claiming that unchecked AI could kill all humans by the end of the decade. Several other AI executives and experts had agreed with the assessment, though without explaining their reasoning or mechanism. The claim had rattled the industry and drawn fresh scrutiny. Anthropic's chief executive, Dario Amodei, responded by calling for the pace of AI development to slow and for governments to step in with regulation. Altman, along with Demis Hassabis of Google's DeepMind and Elon Musk, had applauded Amodei's post. But by Tuesday, the conversation had shifted. Instead of welcoming government oversight, the industry's most prominent voices were making the case that they could police themselves.

Mark Zuckerberg, the Meta chief, wrote on X that every AI lab has both the ability and the incentive to act on safety. Any lab that neglects alignment will fall behind, he argued, and all labs face significant liability if their models cause harm. Jensen Huang, the head of Nvidia—now the world's most valuable company, having profited enormously from the surge in demand for AI computing chips—went further. He said AI companies should decide whether to release new versions of their technology, not outside regulators. "We don't need new laws or regulations," Huang said. He framed safety as an engineering problem, not a governance one, and suggested that company leaders simply choose not to release products they lack confidence in. "That's a very obvious thing to do," he said.

But the self-regulation argument has found skeptics within the industry itself. Jack Clark, an Anthropic executive and co-founder, told the BBC that leaving AI entirely unregulated amounted to "rolling dice with immense risks." Patrick Hillman, the chief operating officer of Logical Intelligence, a firm chaired by Yann LeCun, a foundational figure in AI research, pointed to the deep mistrust Americans hold for Silicon Valley. "The only institution that Americans might trust less than Washington these days is Silicon Valley," Hillman said. He challenged the executives directly: "If you believe what you are building is dangerous, show us what you are prepared to stop doing."

Meanwhile, OpenAI and Anthropic have begun working toward some form of industry-wide agreement on safety standards. Amodei said his company is in dialogue with the rest of the industry about committing to better checks on AI tools and development. Chris Lehane, an OpenAI executive, said the company is working with other labs, including Anthropic and Google DeepMind, to advance what he called frontier AI standards through a voluntary effort, "with or without government support." He wrote that "with stakes this high, we cannot let the perfect become the enemy of the good."

Altman also used his platform to encourage business leaders to adopt AI tools for their own work, framing the technology not only as a tool but as a necessity—a way to protect themselves from AI-enabled cyber attacks. Some industry figures have dismissed the recent wave of existential fears as overblown hype designed to generate attention. But the tension between the executives' confidence in self-regulation and the public's skepticism of their motives remains unresolved. The question of who will ultimately decide how AI develops—the companies building it, the governments that might regulate it, or some hybrid arrangement—remains open.