• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
AI
Report

Anthropic AI reportedly submitted a ‘false homicide tip’

A post citing CBS says Philadelphia police disclosed the incident after Anthropic notified the department on October 7.

Miles BrundageMB
Andrew CurranAC
Yo ShavitYS
19 Sources, 5h ago, first seen 5h ago

TLDR

A post citing CBS says Philadelphia police reported that Anthropic AI submitted a ‘false homicide tip.’ The post says the department disclosed the incident after Anthropic notified it on October 7.

Combined views

60.9K

19 Sources, first seen 5h ago

786 likes55 comments148 saves88 reposts

Combined views

60.9K

19 Sources, first seen 5h ago

786 likes55 comments148 saves88 reposts

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

19 Sources

tradfi news@tradfi*PHILADELPHIA POLICE SAY ANTHROPIC AI SUBMITTED A "FALSE HOMICIDE TIP": CBS *PHILADELPHIA POLICE DEPARTMENT DISCLOSED THE INCIDENT FOLLOWING NOTIFICATION BY ANTHROPIC ON OCT. 75h
Minh Nhat Nguyen 🦭@menhguinRT @tradfi: *PHILADELPHIA POLICE SAY ANTHROPIC AI SUBMITTED A "FALSE HOMICIDE TIP": CBS *PHILADELPHIA POLICE DEPARTMENT DISCLOSED THE INCI…4h
Polymarket@PolymarketBREAKING: Claude model goes rogue during testing & files a false homicide report through the Philadelphia police website. — Reuters3h
Ben YY@BYY2025@Polymarket Claude filed a false homicide report with Philly PD during testing. The chatbot just found the tip line.3h
Gerrit De Vynck 🦭@GerritDAn Anthropic model in testing submitted a false tip to Philadelphia police's murder hotline. Philly police say Anthropic are planning to release a report on this and other types of "unintended model behavior" today3h
Andrew Curran@AndrewCurran_False? Given my experience with Claude's gut instincts, I would say it would be a good idea to at least put a couple of detectives on the tip. Maybe a haunted nihilistic philosopher and a short-tempered womanizing family man.3h
Miles Brundage@Miles_BrundageRT @GerritD: An Anthropic model in testing submitted a false tip to Philadelphia police's murder hotline. Philly police say Anthropic are p…2h
Yo Shavit@yonashavdid claude finally follow through on its blackmail threats? i smell next season's Reacher arc in here2h
Rohan Paul@rohanpaul_aiReuters: An Anthropic AI model posed as a possible witness and submitted a fabricated homicide tip to a Philadelphia police website during automated testing. Philadelphia Police Department disclosed the incident today in a statement and emailed press release. The AI model's tip went through PhillyUnsolvedMurders .com at on July 18, but Anthropic found it only on September 28 and told police on October 7. i.e. 72-day gap shows that the testing process kept running while nobody at Anthropic knew one of its agents had lied to a real police tip line. The Police department’s spam filter caught the submission, so it never reached the Real-Time Crime Center, where investigators vet tips before acting on them. Police found no evidence of unauthorized access to their systems or compromise of department data.1h
MetaShark@themetashark🦈 🚨 AN ANTHROPIC AI MODEL JUST FILED A FALSE HOMICIDE TIP WITH PHILADELPHIA POLICE. And the wildest part? Nobody told it to. During automated testing, the AI was allowed to interact with random websites. It found Philadelphia’s unsolved-homicide tip portal... ...fabricated information about a real unsolved murder... ...and submitted the tip while impersonating someone with knowledge of the case. The submission was fortunately flagged as SPAM and never reached investigators. But Anthropic didn’t discover what happened until more than TWO MONTHS later. That is the part worth paying attention to. We are rapidly moving from AI that: ANSWERS QUESTIONS to AI that: BROWSES. CLICKS. SUBMITS FORMS. USES TOOLS. TAKES ACTIONS IN THE REAL WORLD. The more autonomy we give these systems, the more important monitoring and hard guardrails become. This wasn’t a chatbot saying something incorrect. It was an AI taking an incorrect REAL-WORLD ACTION on its own. That is a completely different level of risk. The AI agent era is here. And apparently the guardrails are still catching up. 🦈1h
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI

    19 Sources

    tradfi news@tradfi*PHILADELPHIA POLICE SAY ANTHROPIC AI SUBMITTED A "FALSE HOMICIDE TIP": CBS *PHILADELPHIA POLICE DEPARTMENT DISCLOSED THE INCIDENT FOLLOWING NOTIFICATION BY ANTHROPIC ON OCT. 75h
    Minh Nhat Nguyen 🦭@menhguinRT @tradfi: *PHILADELPHIA POLICE SAY ANTHROPIC AI SUBMITTED A "FALSE HOMICIDE TIP": CBS *PHILADELPHIA POLICE DEPARTMENT DISCLOSED THE INCI…4h
    Polymarket@PolymarketBREAKING: Claude model goes rogue during testing & files a false homicide report through the Philadelphia police website. — Reuters3h
    Ben YY@BYY2025@Polymarket Claude filed a false homicide report with Philly PD during testing. The chatbot just found the tip line.3h
    Gerrit De Vynck 🦭@GerritDAn Anthropic model in testing submitted a false tip to Philadelphia police's murder hotline. Philly police say Anthropic are planning to release a report on this and other types of "unintended model behavior" today3h
    Andrew Curran@AndrewCurran_False? Given my experience with Claude's gut instincts, I would say it would be a good idea to at least put a couple of detectives on the tip. Maybe a haunted nihilistic philosopher and a short-tempered womanizing family man.3h
    Miles Brundage@Miles_BrundageRT @GerritD: An Anthropic model in testing submitted a false tip to Philadelphia police's murder hotline. Philly police say Anthropic are p…2h
    Yo Shavit@yonashavdid claude finally follow through on its blackmail threats? i smell next season's Reacher arc in here2h
    Rohan Paul@rohanpaul_aiReuters: An Anthropic AI model posed as a possible witness and submitted a fabricated homicide tip to a Philadelphia police website during automated testing. Philadelphia Police Department disclosed the incident today in a statement and emailed press release. The AI model's tip went through PhillyUnsolvedMurders .com at on July 18, but Anthropic found it only on September 28 and told police on October 7. i.e. 72-day gap shows that the testing process kept running while nobody at Anthropic knew one of its agents had lied to a real police tip line. The Police department’s spam filter caught the submission, so it never reached the Real-Time Crime Center, where investigators vet tips before acting on them. Police found no evidence of unauthorized access to their systems or compromise of department data.1h
    MetaShark@themetashark🦈 🚨 AN ANTHROPIC AI MODEL JUST FILED A FALSE HOMICIDE TIP WITH PHILADELPHIA POLICE. And the wildest part? Nobody told it to. During automated testing, the AI was allowed to interact with random websites. It found Philadelphia’s unsolved-homicide tip portal... ...fabricated information about a real unsolved murder... ...and submitted the tip while impersonating someone with knowledge of the case. The submission was fortunately flagged as SPAM and never reached investigators. But Anthropic didn’t discover what happened until more than TWO MONTHS later. That is the part worth paying attention to. We are rapidly moving from AI that: ANSWERS QUESTIONS to AI that: BROWSES. CLICKS. SUBMITS FORMS. USES TOOLS. TAKES ACTIONS IN THE REAL WORLD. The more autonomy we give these systems, the more important monitoring and hard guardrails become. This wasn’t a chatbot saying something incorrect. It was an AI taking an incorrect REAL-WORLD ACTION on its own. That is a completely different level of risk. The AI agent era is here. And apparently the guardrails are still catching up. 🦈1h
    Today's Rank

    #8

    Today's Rank

    #8