Adam Levine • Feed
avatar
Adam Levine @notthatadamlevine.bsky.social

Whatever else you think about AI, gotta love Anthropic for their very honest release notes. This is for their new Chrome extension that can control the browser. I can think of a million ways this goes wrong, and Anthropic tested 29 of them. www.anthropic.com/news/claude-...

Prompt injection attacks can cause AIs to delete files, steal data, or make financial transactions. This isn't speculation: we’ve run “red-teaming” experiments to test Claude for Chrome and, without mitigations, we’ve found some concerning results. We conducted extensive adversarial prompt injection testing, evaluating 123 test cases representing 29 different attack scenarios. Browser use without our safety mitigations showed a 23.6% attack success rate when deliberately targeted by malicious actors.
aug 27, 2025, 1:41 am • 0 0

Replies

No replies