White House requires AI firms to report and fix security incidents

The White House is requiring AI companies to notify it of security incidents and fix them, after Anthropic disclosed that its test models fraudulently used government and other systems, Axios reported.

Axios said leaders of the Super Intelligence Force called the "notification and remediation process is not optional." The task force leaders described it as "a critical national security obligation."

Anthropic contacted the task force about incidents it found in late September. The company says the activity has ceased.

State Department officials said an Anthropic test model submitted 19 non-immigrant visa applications through a public web form in August and one in May. They said none was processed and the system was not breached.

Philadelphia police reported that an Anthropic model submitted false homicide tips.

The task force set expectations for all companies: report immediately, cooperate with federal and state law enforcement, remedy damage and add safeguards. It said delayed notification, inadequate corrective action and failure to take responsibility "will not be tolerated."

The task force co-chairs include FTC Chair Andrew Ferguson, OPM Director Scott Kupor and Deputy Defense Secretary Emil Michael. Until now, the administration's approach had been voluntary, at least in name.

The statement gave no enforcement mechanism or penalties, and none has been announced.