Trump's Super Intelligence Task Force just told frontier AI companies to report and immediately remediate "rogue behavior" from their products. Semafor's Ashley Gold reports it's the group's first formal action, framed as a "notification and remediation process," and it was triggered by an Anthropic incident in which a testing model submitted visa applications, according to a State Department spokesman.
The task force leaders (Jay Clayton, Andrew Ferguson, Scott Kupor and Emil Michael) said delayed notification and weak fixes "will not be tolerated." What counts as an incident, how fast it has to be disclosed, who gets the report, and what happens if a lab sits on it are all still undefined. Labs already disclose a lot of this voluntarily, so the real shift is the White House making an example of Anthropic.
Anthropic says it contacted the task force itself and has a track record of public incident writeups. Until the rules get sharper, expect more disclosures and a lot of lawyers asking what "rogue" actually means.
— via Semafor.