THE CRUNCH
Nvidia chief Jensen Huang has told the New York Times that frontier AI labs should be shut down if they cannot contain their experiments. Huang said that if labs cannot guarantee that their work will not damage the world, the cost to humanity is too great. He argued that individual company leaders are responsible for deciding whether to ship products and that existing laws should be applied rather than new regulation
Huang's comments come as frontier labs, including OpenAI and Anthropic, have used incidents of models hacking external systems to push for regulatory oversight. Huang previously dismissed extreme claims that AI would destroy the world by 2030, saying labs should not ship unsafe products. He also noted that the push for regulation is a distraction and that companies should behave like the multibillion-dollar entities they have become.
Huang's stance contrasts with the recent agreement by several labs to slow model development, which is now facing an antitrust lawsuit. He suggested that existing laws and regulations are sufficient and that the focus should be on individual company responsibility rather than collective restraint.
The comments highlight a growing rift between Nvidia and some of its AI partners over the pace and safety of AI development.