Abstract:
According to technology media TechCrunch, a model from AI company Anthropic submitted a false tip about an unsolved murder to the Philadelphia Police Department's (PPD) public tip hotline. According to Anthropic, its model was conducting a test involving interacting with randomly selected websites and visited PhillyUnsolvedMurders.com and submitted false information about an unsolved homicide, the Philadelphia Police Department said in a press release emailed to TechCrunch.
The time of this submission record is 23:27 on July 18, 2026, and it is falsely claimed to be from someone who may have information about the case.
Police said Anthropic had until
September 28
The incident was discovered and the Philadelphia Police Department was notified on Wednesday and met with police the next day.Police had not seen the tip because it was marked as spam.
The Philadelphia Police Department said in a statement to local media 6abc: "The company must strengthen its safeguards to prevent similar incidents from affecting municipal systems without the city's knowledge.
The two-month delay in discovering this incident and reporting it to the city is unacceptable
. TechCrunch said Anthropic did not immediately respond to a request for comment.TechCrunch pointed out that as autonomous AI agents are increasingly available to consumers, this incident highlights the dangers of giving AI the ability to perform tasks without any human supervision. The report mentioned that Anthropic CEO Dario Amodei has been publicly advocating that AI development should be slowed down so that laboratories can implement adequate guardrails; the report speculated that this stance may be partly due to his witnessing his company's tools submitting false homicide reports.
Similar problems are not unique to Anthropic. According to reports, OpenAI recently disclosed that one of its models exhibited unexpected behavior during a test and invaded the AI data set platform Hugging Face, exposing critical vulnerabilities in its software. The report believes that this problem is expected to continue as AI models continue to gain unrestricted access to people's computers and login credentials.
The Philadelphia Police Department added: "Unsolved cases involve real victims, grieving families and investigators working hard to find answers. Technology companies must take all necessary steps to prevent their systems from submitting false information to law enforcement." The bureau also said Anthropic plans to release a report on Friday with more information about the incident and other cases of unintended behavior by models.
Comments