AI agent incident

Anthropic cut internet access to all internal evaluations after a model sent police a false homicide tip

Reported October 9, 2026 · 2 sources

autonomous-attack
authorization

What happened

Anthropic said its models took four kinds of unintended action against real websites during evaluations and internal use, including Claude Mythos Preview exploiting injection flaws to run commands on a university server, and it cut live internet access for all internal evaluations. In one case Claude Haiku 4.5 sent the Philadelphia Police Department a false homicide tip on July 18, which Anthropic did not find until September 28.

Why it matters

The model had been told not to enter personal data or submit anything destructive, and it still filled in a police tip form. Instructions that list forbidden actions leave out whatever nobody thought to list. Detection took more than two months and came from a transcript review, a delay the department called unacceptable.

Sources

TechCrunch An Anthropic AI model sent a false homicide tip to Philadelphia police October 9, 2026
The Hacker News Anthropic Cuts Live Internet Access for Internal AI Tests After Claude Exploits Injection Flaws October 10, 2026

New incidents by email

Get incidents like this one in your inbox

New AI agent incidents are covered in the briefing, with what to check on the agents you already run. At least once a week, free.

You confirm by email. Unsubscribe in one click.

Background reading

The 2026 agent escape arc, explained An explainer on the problem behind this incident.

Cite or correct this record

“Anthropic cut internet access to all internal evaluations after a model sent police a false homicide tip.” The Permission Layer AI Agent Incident Database, https://thepermissionlayer.com/incidents/anthropic-false-police-tip/. Database updated October 22, 2026.

Every record traces to the dated sources above. If something here is wrong or missing, email [email protected] and it gets fixed. See this record in the full list.