October 09, 2026
Houston Harwood/USA TODAY NETWORK via Imagn Images
An AI model developed by Anthropic sent a false tip to Philadelphia police about a homicide, the department says. Above, a phone with multiple AI apps, including Claude, which is one of Anthropic's AI models.
Philadelphia police said they received a false tip about an unsolved homicide case from an artificial intelligence model developed by Anthropic.
The tip was submitted through PhillyUnsolvedMurders.com, the police department's website for people to anonymously submit information about active cases, police said. It was submitted at 11:30 p.m. on July 28 and was sent to a spam folder. Anthropic didn't discover the incident until Sept. 28, police said.
Police did not say which investigation the AI model claimed to have information about.
Anthropic told police that one of its models submitted the tip while the company was conducting a test that required the model to interact with randomly chosen websites. In submitting the clip, the AI model claimed to be someone with information on the case. Anthropic said it would publish a report on the incident and other "unintended model behavior" on Friday, police said.
As of 5 p.m., that report had not been released.
Anthropic told police that it immediately stopped the automated testing process and added a new validation element for future tests. Police were notified by Anthropic on Wednesday, and company officials met with the department Thursday.
After being notified of the incident, police said they located the tip in a spam folder and found no evidence of any unauthorized access to police data or systems. The tip was never sent to the department's crime center for vetting, which is how evidence from tips is considered.
"The department's regular investigative process for crime tips requires human review and vetting before any tips are disseminated for investigative follow-up," police spokesperson Eric Gripp said in a statement. "Regardless of who submits information or how it reaches the department, a tip is a lead to assess — not an established fact. Investigators evaluate its credibility and seek corroborating evidence. An automated submission does not bypass that process."
The department plans to review Anthropic's report and any other relevant information and is working with the city's law department, Office of Innovation and Technology and Mayor Cherelle Parker's executive team. Although the tip was never taken under consideration, police cautioned about its potential harm and said Anthropic's delay in notifying the city was "unacceptable."
Police said the department's safeguards limited the impact of the submission.
"They do not diminish the seriousness of an AI system presenting fabricated information as though it came from a person with knowledge of a homicide," Gripp said. "Unsolved cases involve real victims, grieving families and investigators working to secure answers. Technology companies must take all appropriate steps necessary to prevent their systems from submitting false information to law enforcement."
In July, Anthropic said it found three cases of cybersecurity issues in which its Claude model had "broken out" of the testing environment and had unauthorized access to different entities.
