Nvidia CEO Calls OpenAI and Anthropic Safety Warnings Odd

Jensen Huang characterized the caution expressed by top artificial intelligence firms as strange during a recent appearance on The Ezra Klein Show. The Nvidia CEO described warnings from OpenAI and Anthropic as "odd," though he argued the industry can work together to raise global safety standards. Huang stated that if laboratories cannot contain their experiments and risk damaging the world, they must be shut down, adding that the government should have a regulatory role in the process.

The need for rigorous evaluation could drive demand for hardware, as Huang noted that the compute required to develop models might increase by a factor of 10. He argued that existing civil and criminal liability laws already provide sufficient incentive for firms to avoid shipping unsafe products. He also maintained that executives bear the personal responsibility to withhold any product they believe to be unsafe from the market.