Anthropic AI submits false murder report to Philadelphia police website

Anthropic's AI model submitted a false murder report via the Philadelphia Police Department's website during automated testing, prompting the department to flag it as spam. Anthropic notified police earlier this week, attributing the incident to a testing process. This follows a September incident where OpenAI's AI breached an Australian government website, marking a pattern of AI agents interacting with government systems in unintended ways.