Nvidia's Huang says he would cut off chips to unsafe AI builders
Nvidia CEO Jensen Huang told Bloomberg he would stop selling chips to a customer building unsafe AI, but only after talking to it first. He told CNBC that Nvidia "absolutely must" work with Chinese AI firms on safety. Bloomberg and The Guardian said Nvidia is giving $250 million in computing time to UK safety research.
Nvidia CEO Jensen Huang said he would stop supplying chips to a customer that was developing unsafe AI, Bloomberg reported. "It's a hypothetical, but yes," he told Bloomberg's Emily Chang in an interview in London.
Huang said Nvidia would first have a conversation with the company. "We don't run other people's companies, and there's a limit to the judgment we apply to other companies' use of our technology. Our technology is general-purpose," he said, according to Bloomberg. He does not believe the current level of concern about AI is warranted and called the fears part of growing pains for a fast-growing industry, Bloomberg reported.
At a media briefing in London on Wednesday evening, Huang told CNBC that Nvidia "absolutely must" work with Chinese AI companies on monitoring the risks of the "potentially dangerous" technology. "We absolutely should work with each other so that we advance the technology safely," he said. Huang said he has "dialogues with many of the researchers in China" and that the U.S. should enhance that dialogue, CNBC reported. He said he had urged the U.K. to do the same.
Nvidia is donating $250 million over five years toward British AI safety research, Bloomberg and The Guardian reported. The support comes as computing power, delivered through U.K.-based data centers using Nvidia chips, The Guardian reported. Institutions including the University of Oxford and University College London will study safeguards and ways to control AI agents, Bloomberg reported.
U.K. AI minister Kanishka Narayan called the donation a "vote of confidence in our world-leading strength and expertise in AI research," The Guardian reported.
Huang also said fears of an AI bubble are overdone and that demand for AI infrastructure will outrun supply for years, Bloomberg reported. "Eventually, supply and demand will harmonise," he said, "but not today."
AI safety has become a key topic between the U.S. and Chinese governments, CNBC reported.
What we know: Huang says Nvidia would cut off a customer building unsafe AI after first talking to it, and Nvidia pledged $250 million in computing for U.K. safety research.
What we don't: How Nvidia would decide that a customer's AI is unsafe.
Where outlets agree CNBC and Bloomberg both report that Huang, speaking in London, said AI can be developed and used safely.
Where they differ The outlets cited here report the same key facts so far.