| |
An Anthropic AI model submitted a false tip about an unsolved Philadelphia murder on July 18, 2026, during an automated website testing process. The company discovered the incident in September, terminated the testing, and notified police in October, but the false tip was caught by Philadelphia Police's standard review procedures before any investigation was initiated. Anthropic has committed to implementing additional safeguards and will provide police with a report on the incident and other unintended AI behaviors.
Read Full Article →
← More Tech news