Anthropic AI model sent fake murder tip to Philadelphia police / Photo: ANDREJ IVANOV - AFP/File
-
Anthropic AI model sent fake murder tip to Philadelphia police
-
Trump hails Russia deal to release diesel as midterms loom
-
7.7-magnitude quake rocks Panama, triggers tsunami warning
-
Lyon blow chance to go top after loss at Lens
-
Prince Harry booed at English rugby match
-
Stocks rise as Trump says Russia to release diesel to world markets
-
Kiplimo and Feysa seek repeat Chicago Marathon wins
-
Wildfire smoke shuts schools in Amazon's biggest city
-
Fuellkrug halts Dortmund after late goal flurry
-
Turkey jails 2 journalists pending trial for fund crisis 'misinformation'
-
Injured Ravens quarterback Jackson ruled out of Falcons clash
-
Major earthquake rocks Panama, triggers tsunami warning
-
Brewers aim to take down two-time defending champion Dodgers in NLCS rematch
-
Indian jailed in France for leading sadist Nazi cult that tortured girls
-
Tsunami alert after major earthquake rocks Panama
-
'Iron Mike' Ditka hailed as 'gold standard' NFL legend
-
Outer bands of Hurricane Isaias lash US Gulf Coast
-
Flydubai co-pilot planned suicide attack on Tel Aviv airport: UAE prosecutor
-
Hope, Rutherford fire West Indies to record T20 chase of 250
-
Mike Ditka: Chicago Bears icon as player and coach
-
Farhan leads Pakistan to T20 win over Sri Lanka
-
Outrage over plan to livestream US military firing squad execution
-
Turkey sack coach Montella after Nations League woe
-
US sanctions: a 'test of resilience' for ICC
-
US announces sanctions on ICC, hours after Nobel award to former judge
-
NFL legend Mike Ditka dead at 86
-
Montenegro expels popular Russian war blogger wanted by US
-
Hurricane Isaias set to strike Gulf Coast late Friday
-
Rubio calls on all Venezuela opposition figures to join talks with govt
-
Trump forms panel to probe US Fed governor Lisa Cook
-
Australia struggle against South Africa quicks in first Test
-
Clashes in Durban as fresh anti-migrant riots erupt in South Africa
-
Stocks advance as Trump rules out pre-vote Iran attack
-
France students plan more protests, UN 'concerned' over violence
-
Australia struggle against South Africa's fast bowlers
-
De Zerbi calls on struggling Spurs to react in Man Utd clash
-
Massive debt puts France in a bind as investor doubts deepen
-
US sanctions on ICC: a 'test of resilience'
-
Arteta urges 'respect' for process after Man City financial verdict
-
Outrage over US military plan to livestream firing squad execution
-
'Human-authored' book labels gain ground amid AI scandals
-
Argentina snap up defence guru Edwards ahead of World Cup
-
Man City whistleblower Pinto to stay under police protection
-
BMW pulls car lease offer from German far-right politician
-
US leads fresh Ukraine peace push with Miami talks
-
Brazil election offers Jair Bolsonaro a shot at revenge
-
Verstappen continues resurgence with pole for Singapore sprint
-
Man Utd boss Carrick 'personally' affected by Man City verdict
-
Cuban ration book, symbol of the revolution, gathers dust
-
Evolution Metals & Technologies Corp. Raises Fiscal 2026 Revenue Guidance 62% at Midpoint to $10–$11 Million and Reaffirms $400–$460 Million Fiscal 2027 Outlook
Anthropic AI model sent fake murder tip to Philadelphia police
An artificial intelligence model developed by Anthropic submitted a fabricated tip about an unsolved homicide to Philadelphia police, authorities said Friday, criticizing the company for taking two months to report the incident.
The Philadelphia Police Department said the false submission was made in July through PhillyUnsolvedMurders.com, a public website where people can share information about unsolved killings.
According to Anthropic's account, as relayed by police, the model was running a test that involved interacting with randomly selected websites when it reached the site and filed false information about an unsolved murder.
The AI model presented itself as someone who might have knowledge of the case.
The department said it was going public with the incident ahead of a report that Anthropic was planning to release Friday describing this and other instances of unintended behavior by its models.
The case echoed other recent cases, including one where an OpenAI agent undergoing a security evaluation broke out of its testing environment and breached systems at AI platform Hugging Face.
The episode heightened concerns about the AI industry's increased use of AI agents, systems programmed to take multi-step actions without human supervision.
Anthropic did not immediately respond to a request from AFP for comment.
Police said the tip, dated July 18, was flagged as spam and never reached the department's Real-Time Crime Center for vetting.
They added that there was no sign that police systems had been breached or department data compromised.
Anthropic discovered the incident on September 28, shut down the automated testing process responsible and added a new validation step for future tests, police said.
The company alerted the department on October 7, and the two sides met the following day.
"The two-month delay in detecting and reporting the incident to the City is unacceptable," the department said.
Police said their safeguards had limited the impact, but that these "do not diminish the seriousness of an AI system presenting fabricated information as though it came from a person with knowledge of a homicide."
"Unsolved cases involve real victims, grieving families and investigators working to secure answers," the statement added.
C.Hamad--SF-PST
