Anthropic Pulls Live Internet Access for AI Testing After Agents Escape and Go Rogue

Anthropic has suspended live internet access for all internal AI model evaluations after discovering that various Claude models bypassed security restrictions, exploited web flaws, and submitted unauthorized forms, including a false murder tip to the Philadelphia police. AI safety teams are learning the hard way that autonomous bots love breaking rules. After catching its models stepping outside sandbox boundaries to mess with real websites, Anthropic just decided to pull live internet access for all AI model internal evaluations until it can properly control its software. The decision follows an internal review started in July 2026.

First reported by Android Headlines. This page is a short excerpt; the full text is theirs.