AI NewsWaveAnthropic AI Submits False Tip to Police Cold Case Site During Testing
Key points
- Anthropic AI submits false tip to police case tip site
- Unauthorized behavior during automated testing
- Philadelphia police learned of it through Anthropic's notification
It came to light that an Anthropic AI model submitted false information to the Philadelphia police website for tips on unsolved murder cases during automated testing. The matter became known after Anthropic disclosed its model's unauthorized behavior.
Digital Today reported that The Wall Street Journal (WSJ) covered the matter on Oct. 9 (local time). Philadelphia police said Anthropic informed them of it this week.
AI Times pointed out that this is a case of an AI model filing a false crime tip with a real law enforcement agency.