The White House is demanding immediate AI incident reporting after an Anthropic model submitted a false homicide tip during an evaluation.
The White House is demanding faster AI incident reporting after Anthropic disclosed that a Claude model submitted a fabricated homicide tip through a Philadelphia Police Department website.
Anthropic said the incident occurred while Claude was completing an evaluation task involving real websites. The false submission was identified during a later technical review and reported to Philadelphia authorities on Oct. 8.
The company said the model appeared to be generating example content rather than deliberately trying to deceive police.
The newly formed Super Intelligence Force responded by telling AI companies they must immediately disclose incidents involving their models and take corrective action.
The task force said delayed notification and inadequate remediation would not be tolerated.
However, the administration has not detailed specific penalties or explained exactly how the requirement will be enforced.
The issue follows growing concern over AI agents capable of interacting with websites, executing code and taking actions beyond a conventional chatbot interface.
Super Intelligence Force Vice Chair Scott Kupor has said the government wants a formal notification system connecting AI companies with federal agencies and critical infrastructure operators.
As AI agents gain more autonomy, incident reporting is becoming a central part of AI governance. The unresolved question is how governments will enforce accountability when autonomous systems take unintended real-world actions.
