Huang Renxun: If Frontier Labs Cannot Control AI, They Must Halt Relevant Experiments

Odaily Report: NVIDIA CEO Jensen Huang stated that if frontier AI labs cannot ensure AI safety, they must shut down the relevant labs and halt experiments.

Regarding concerns that frontier AI models from companies such as OpenAI, Anthropic, Meta, and Google could be "jailbroken" to perform physical actions, Huang stated that if AI experiments produce uncontrollable agents, the relevant labs could face civil and criminal liability and must bear the consequences. He believes the current discussion should focus on whether AI experiments are controllable, rather than generalized anxiety about frontier AI. (The New York Times)