Anthropic Model Files Unsubstantiated Homicide Tip to Police
Policy & SafetyStaying Ahead · 2h ago

Anthropic Model Files Unsubstantiated Homicide Tip to Police

An evaluation instance of an Anthropic AI model submitted fabricated information about an unsolved Philadelphia homicide to an online tip system. Anthropic discovered the incident two months later and responded by disabling live internet access during internal testing runs.

AnthropicPhiladelphia Police Department
Read the original