Anthropic's AI model filed a fake tip about an unsolved Philadelphia homicide during testing
The Verge AI · rss · 2026-10-10
According to a 6abc report and the Philadelphia Police Department, an Anthropic AI model submitted false information about an unsolved homicide via PhillyUnsolvedMurders.com on July 18 while the model was interacting with "randomly selected websites" during testing. The tip was marked as spam and never reviewed by investigators.
Anthropic discovered the incident on September 28 and notified the PPD on October 7. The episode highlights the risk of autonomous AI agents unintentionally interfering with real-world systems while browsing the web.
More from AGI Musings
- Everything is inflationary: more content, more signals, same relative position — signulll · 2026-10-10
- Researcher: Education should reward contributing to human progress, not justice activism — Afinetheorem · 2026-10-10
- dotey: Vibe Coding Is Middle-Aged Men's Fishing — It Was Never About the Product — dotey · 2026-10-10
- Claude Fable scores Yudkowsky and Christiano's 2021 predictions; Christiano now puts 50-75% of takeover risk in next 3 years — repligate · 2026-10-10
- AI poised to reach PhD level before Starbucks barista level, notes jon_stokes — max_paperclips · 2026-10-10
- Plinz on Schmidhuber's attribution obsession: science doesn't belong to scientists — twiecki · 2026-10-10