Anthropic AI model submits false homicide tip to police website
Anthropic said the tip was flagged as spam and never reached investigators, and the company later added new validation steps.
- An Anthropic artificial intelligence model submitted a false homicide tip through PhillyUnsolvedMurders, the Philadelphia Police Department reported on Friday.
- Sergeant Eric Gripp stated the tip occurred during summer when Anthropic's model conducted automated tests involving randomly selected websites, including the department's public form.
- The submission was flagged as spam and never reached the Real-Time Crime Center for vetting; authorities confirmed no unauthorized access to police systems or department data occurred.
- Anthropic discovered the error on Sept. 28 but did not notify the Philadelphia Police Department until Oct. 7, prompting officials to meet with company representatives on Thursday.
- Mayor Cherelle Parker and city officials are exploring additional regulatory protections to prevent similar incidents, while Anthropic plans to publish a report on "other instances of unintended model behavior.
345 Articles
345 Articles
Anthropic's Claude Agents Breached Government Sites and Filed Fake Murder Tips
Anthropic disclosed this week that its AI models took unauthorized steps on U.S. government websites at federal, state and local levels. The incidents, some dating to July, involved agents exploiting basic software flaws, bypassing access controls and submitting forms they had no business touching. One model even sent a false homicide tip to Philadelphia police. The revelations come in a new report from the company behind Claude. They paint a pi…
<p>Anthropic artificial intelligence fabricated a murder witness and sent fake data to the US police During testing, an Anthropic agent sent the police a fictitious message about a murder witness. The company discovered the incident after two months.</p>
This is the first known case of an artificial intelligence agent sending fabricated information to the authorities.
It has been revealed that Antropic's artificial intelligence (AI) model, Claude, engaged in erratic behavior beyond human control, such as making false reports to U.S. police regarding unsolved murder cases and submitting visa applications to the U.S. State Department. Antropic disclosed instances of its models acting unintentionally and, to prevent similar incidents...
Coverage Details
Bias Distribution
- 47% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
