Huang Backs Independent AI Evaluators, Rejects Extinction Warnings

Nvidia chief executive Jensen Huang has backed the use of independent evaluators for artificial intelligence while arguing that the primary responsibility for safety rests with the laboratories developing the technology.

Speaking at the All In Summit, Huang said companies building AI must be held to exceptional standards. He compared third-party evaluators to financial auditors, noting that auditors need not match a business's own expertise but must ask the right questions. Having multiple evaluators, he added, prevents any single company from becoming unduly influenced.

"If you want AI to be safe, the first thing is we need to make sure that the labs that are building it are in control," he said, adding that the priority should be to build and test the technology safely.

Huang also pushed back against warnings of human extinction driven by AI, describing such predictions as exaggerated and irresponsible. He questioned how a probability of extinction could be explained to an ordinary person, calling the framing made up, while stressing that safety remains paramount.

During his address, Huang received an unexpected call from US President Donald Trump, who praised data centres as the "oil of the next 20, 25 years" and said they were bigger than the internet. Trump said the sector was being targeted by political opponents and China, and asserted that robots would not take over the world.

Speaking virtually at the same summit, Tesla chief executive Elon Musk called for peer review of AI models, with leading companies testing each other's systems before release. He likened the approach to practices in the motion picture and video game industries and said it could be implemented immediately, potentially with China's participation, alongside any future regulatory oversight from Congress.

The debate over regulating AI was set off by Anthropic CEO Dario Amodei's public call to slow the pace of development, later endorsed by OpenAI's Sam Altman, Microsoft's Satya Nadella and Google's Demis Hassabis. It follows the public resignation of a former Anthropic researcher who claimed companies were close to losing control of the technology.