The White House’s plan to vet potentially dangerous AI is cloaked in secrecy

The Trump administration has finalized a framework for testing new AI models for safety and cybersecurity risks, according to a report from The Guardian. The plan to vet potentially dangerous AI was completed this week. The framework follows months of talks with technology industry leaders. The White House does not plan to release its policy publicly, the Guardian reports. The article is headlined 'The White House's plan to vet potentially dangerous AI is cloaked in secrecy'.

Under the finalized framework, testing criteria will be shared only with a select few technology companies, according to the Guardian. The report does not name those companies. It also does not provide details of the testing standards themselves. The report does not specify which AI models would be subject to the testing or what thresholds would trigger a review. The White House has not announced whether or when the framework might be made public.

Keeping the testing criteria private means the standards will not be subject to public scrutiny. Selective sharing with a limited number of companies could create differences in knowledge about how AI models are evaluated. The report frames the confidentiality as a defining element of the policy. The Guardian notes that the White House plans to keep the framework out of public view. No indication is given of when, if ever, the full policy will be released.