Nvidia boss says AI 'doesn't need new laws' as safety concerns grow

Jensen Huang is pushing back hard against the growing calls for tighter AI rules. Speaking at a Salesforce conference in San Francisco on Tuesday, the Nvidia boss said company leaders should decide whether new AI systems are safe enough to ship, and he rejected what he called a “false choice” between moving fast and protecting people.

His message cuts against a nervous mood building inside the industry. Over the past week, an Anthropic researcher’s viral exit post claiming AI could kill all humans by the end of the decade set off a fresh round of alarm. Anthropic chief executive Dario Amodei then urged governments to regulate the sector and slow development, while OpenAI’s Sam Altman, DeepMind’s Demis Hassabis and Elon Musk publicly backed the warning.

Huang was not having it. He called the current wave of AI a “new industrial revolution” but said safety is an engineering problem, not a reason for lawmakers to rush in. If a company leader does not trust a product, Huang said, the obvious move is to hold it back. “Run as fast as you can,” he said, “but if at any time you feel the institution is not in control, take a pause.”

That self-regulation pitch is exactly what some rivals distrust. Jack Clark of Anthropic told the BBC that turning AI into a “totally unregulated industry” would be rolling dice with huge risks. Patrick Hillman of Logical Intelligence went further, saying tech companies have not earned much public trust and that anyone who thinks their work is dangerous should be ready to stop doing parts of it.

Still, not everyone in the room is waiting for governments to save the day. Amodei said Anthropic is already talking with other AI labs about better safety standards and checks, and OpenAI’s Chris Lehane said the company is working with Anthropic and Google DeepMind on voluntary frontier AI standards. The industry wants a pact. The fight is over whether that should be enough.