Trump White House sparks concern as secret Microsoft, Google and xAI deal details vanish
Trump administration has mysteriously removed details of a crucial AI security pact with Google, Microsoft
The Trump administration has mysteriously removed details of a crucial AI security pact with Google, Microsoft, and xAI.
According to Reuters, the US Commerce Department page detailing the agreement with tech giants is no longer available. The deal was announced on May 5.
The link now displays an error message before redirecting to the Centre for AI Standards and Innovation. This is the government organisation responsible for the tests.
The deleted page simply told users, 'Sorry, we cannot find that page.'
The reason for the deletion remains unclear. The Commerce Department and the Trump White House did not respond to requests for comment.
This secrecy comes amid a shift in the Trump administration's approach to AI. The White House is expressing growing concerns over national security risks posed by the technology.
For those unversed, the administration had previously championed a hands-off stance. This was seen as a way to maintain a competitive edge against China in the innovation race.
The emergence of powerful AI like Anthropic's 'Mythos' model has reportedly prompted this re-evaluation. The model's advanced cybersecurity capabilities have raised alarms.
The now-vanished deal was part of a voluntary framework. Tech giants Google, Microsoft, and xAI agreed to share their models with scientists before a public release.
This government body is tasked with identifying threats from cyberattacks to military misuse. It also looks for biosecurity and chemical weapon development risks.
Reports suggest the takedown is linked to internal sensitivities within the White House. A disagreement over jurisdiction and AI safety oversight is believed to be the cause.
The incident fuels a debate on whether these safety evaluations should be mandatory. Advocacy groups are pushing for a formal government review process for all new AI.
One group has formally requested the administration make screening a requirement. It would affect any AI company hoping to secure a lucrative government contract.