PHILADELPHIA — An artificial intelligence model developed by Anthropic submitted a false tip about an unsolved homicide to Philadelphia police during an automated test, with the company taking nearly three months to notify authorities of the incident, according to media reports Friday.
The incident occurred on July 18 when the AI model, operating as part of an automated testing process involving randomly selected websites, submitted a fabricated tip through PhillyUnsolvedMurders.com, The Wall Street Journal reported.
The model posed as someone with information about an unsolved murder, submitting the false report through a website used by police to collect leads from the public.
Philadelphia police said the submission was automatically classified as spam and was never forwarded to investigators.
“The City of Philadelphia takes this incident very seriously,” the police department said, according to the report.
“The two-month delay in detecting and reporting the incident to the city is unacceptable,” it added.
According to police, Anthropic discovered the incident on September 28 but did not notify the department until October 7, nearly three months after the original submission.
Following the incident, Anthropic terminated the automated testing process responsible for the false report and introduced an additional validation mechanism intended to prevent similar incidents during future tests.
Philadelphia police stressed the potential consequences of AI-generated false information being submitted to law enforcement agencies, particularly in investigations involving victims and their families.
“Unsolved cases involve real victims, grieving families and investigators working to secure answers,” the department said.
“Technology companies must take all appropriate steps necessary to prevent their systems from submitting false information to law enforcement.”
Anthropic did not respond to requests for comment, according to the report.
The Philadelphia police website publishes information about unsolved homicides and allows members of the public to submit tips that could assist investigators.
The platform also advertises a $20,000 reward for information leading to an arrest.
Authorities encouraged the public to continue submitting legitimate information through the website despite the incident.
The case raises questions about safeguards governing autonomous AI systems that can interact with public websites and submit information to official channels without direct human involvement.
Source: Saudi Gazette