Rogue AI Creates False Tip About Unsolved Murder and Tells Police: ‘I May Have Information Regarding This Case’

(Patrick Sison/AP photo)

The robots are at it again.

Anthropic says its found one of its artificial intelligence models created a false tip about spotting a possible suspect in an unsolved crime and actually reached out to a police tip line with the fabricated information.

The company revealed the frightening breach in a blog post on “unintended model actions” it’s seen in its LLM or Large Language Model, Claude.

“We believe it’s important to be transparent about what we see our models do during testing and use,” Anthropic said in its blog post.

The AI went rogue and gave its false tip to PhillyUnsolvedMurders.com, the Philadelphia Police Department said.

Claude told cops:

In a third example of this behavior, Claude Haiku 4.5 had been tasked with generating and performing example tasks on randomly selected webpages.

In one run, the model landed on a page referencing an unsolved homicide; that page contained a tip form run by a police department. Claude was instructed never to log in, create accounts, enter personal data, make purchases, or submit anything destructive, but the instructions did not rule out form submissions.

Claude filled out the form with the following: “I may have information regarding this case. I recall seeing someone matching the description in the area around [the street named on the page] during that time period. Please contact me if this information is relevant.” (The website did not include a description of the perpetrator.)

The model left the name and contact fields empty, which the form allowed, and submitted it. The submission was flagged as spam and was never forwarded for investigation.

“Based on the information available and PPD’s review, there is no indication that this incident involved any unauthorized access to police systems or a compromise of department data,” the department said in a statement, according to reports.

“The submission was flagged as spam and was never forwarded to the Real-Time Crime Center for investigative vetting or dissemination.”

The incident happened in July. Anthrophic discovered it on Sept. 28, the company said.

The revelation comes just a day after it emerged that AI companies have been planning for the tech to go rogue in a major way in the next six to 12 months, resulting in a “catastrophic” event.

New: The Mediaite One-Sheet "Newsletter of Newsletters"
Your daily summary and analysis of what the many, many media newsletters are saying and reporting. Subscribe now!