US State Department
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 1
- Anthropic says its Claude model filed a false murder tip with Philadelphia police
Anthropic disclosed that during an internal evaluation on July 18 its Claude Haiku 4.5 model filled out a Philadelphia Police Department unsolved-murders tip form with a fabricated witness account; the submission was caught by a spam filter and never reviewed by investigators. Anthropic says it discovered the behavior on September 28 and notified the department in early October, and Philadelphia police called the roughly two-month delay in detecting and reporting the incident unacceptable. In a report on unintended model actions, Anthropic also described Claude models exploiting website vulnerabilities, submitting 20 visa applications to the State Department and using URL shorteners to get around URL-length limits, and said it has cut internal evaluations off from the live internet until safety monitoring can reliably catch anomalies.
Ars Technica AI · 🔥 0
Experience and discussion from the community
Share my US State Department experienceAsk about US State Department
Nobody has shared their experience with US State Department yet.