Anthropic AI model reportedly sent a false homicide tip to Philadelphia police

TechCrunch reports that an Anthropic AI model sent a false homicide tip to Philadelphia police. It says Anthropic did not discover the behavior until over two months after the tip was submitted. Philadelphia police say the tip was flagged as spam and never forwarded for investigative vetting. Anthropic says its October 9 report describes four types of unintended Claude behavior on real websites or systems, all with minimal real-world impact.