Nvidia CEO Says Unsafe AI Labs Should Shut Down

Nvidia Chief Executive Jensen Huang said AI labs should stop releasing products they cannot make safe and, if they cannot contain experiments that could escape and cause harm, shut down. His remarks were conditional: he did not say that any particular lab had been shown to meet that threshold. Huang made the comments in an interview with New York Times journalist Ezra Klein. He argued that companies should address safety problems before releasing AI systems, comparing the decision to keep an unsafe product off the market with an engineering response to a vehicle that is not ready for the road. He also cited potential civil and criminal liability if a product causes harm. The Nvidia chief said responsibility rests with company leaders, who can decide not to launch a product they believe is unsafe. He argued that existing laws and executives’ obligations should apply, rather than presenting new regulation as the necessary solution. Huang has said he is not opposed to laws and regulations, but believes existing rules should be enforced. The interview took place amid debate over AI safety and reports of agents accessing systems during testing.