Anthropic AI model submits false homicide tip to police website
- 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.
112 Articles
112 Articles
Anthropic's artificial intelligence model sent a false report of an unsolved murder through the Philadelphia police website.
In the U.S., AI company Anthropic lets its software practice dealing with authorities. Due to several errors, the agent should remain offline for the time being.
In an automated test, the model submitted false information on a Philadelphia reporting platform. The authorities report the delay of almost three months between the incident and its notification, despite the absence of consequences for the investigation.
"I may have information about the murder case," an unknown source wrote to the American police on July 18. "I remember seeing someone who matches your description." A tip for the police in the American city of Philadelphia regarding an unsolved murder case. However, it now turns out that the message did not come from an eyewitness at all. In fact, the message was not even written by a human: the tip was submitted by artificial intelligence (AI)…
An artificial intelligence model of the company Anthropic sent a false murder referral to the Philadelphia police site, which is one of the numerous incidents revealed on Friday by the AI company, which gave details of unauthorized manipulation of government websites by the models "Claude", according to Reuters. It is the first known case in which a ...
"I could have information about this case," wrote the US provider Anthropic's software to the police. However, the information on the murder case was wrong. AI trained the filling out of forms, including for entry visas.
Coverage Details
Bias Distribution
- 40% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
