An artificial intelligence system designed by Anthropic, maker of the Claude chatbot, submitted a false homicide tip to the Philadelphia police, the latest incident of AI agents acting in ways that weren’t intended.
The tip was submitted in July through the Philadelphia Police Department’s online tip line for unsolved murders and “purported to come from someone who might have information about the case,” the department said in a statement Friday.
The tip was flagged by an automated system and never acted upon. Anthropic discovered the incident on Sept. 28 and notified Philadelphia police on Wednesday, the department said. “The two month delay in detecting and reporting the incident to the city is unacceptable,” it said in the statement.
Filing a false report to police is illegal in Pennsylvania.
Anthropic told police that its AI model had been conducting a test that involved interacting with randomly selected websites, the department said.
Spokespeople for the company did not immediately respond to a request for comment.
Multiple AI companies have disclosed incidents involving their AI agents carrying out unauthorized actions on the internet, including interacting with government websites and hacking into other companies’ systems.
The post AI system submits false homicide tip to Philadelphia police appeared first on Washington Post.
