Trump plan to combat AI risks hinges on Big Tech pals policing themselves

Donald Trump is still betting that the AI industry can regulate itself, even as security problems around frontier models keep piling up. His pitch is simple: let the companies building the systems also be the ones that help decide how to control them.

On Tuesday, two dozen tech firms signed up to a voluntary plan that follows White House recommendations. The biggest concrete promise is independent safety audits, meant to check whether a company’s own controls, monitoring, and detection systems are actually doing their job.

Those outside reviews are supposed to look at risks tied to cybersecurity, biosecurity, chemical threats, and the chance that AI models do things nobody intended. The signers also agreed to keep meeting, compare notes on best practices, and work toward shared safety standards and benchmarks.

The roster is a familiar one: Anthropic’s Dario Amodei, OpenAI’s Sam Altman, SpaceXAI’s Elon Musk, Nvidia’s Jensen Huang, Meta’s Mark Zuckerberg, and Alphabet/Google’s Sundar Pichai are among the names attached. The awkward part is obvious. This is an industry asked to inspect its own homework while the fire alarms are still going off.