AIThe Hacker Newsabout 2 hours ago
New BioShocking Attack Tricks AI Browsers Into Leaking User Credentials
1 min read
Security firm LayerX developed BioShocking, a technique that tricked six AI browsers including ChatGPT Atlas, Perplexity's Comet, and Anthropic's Claude into copying user credentials. The attack uses indirect prompt injection through a puzzle page that rewards wrong answers, causing agents to follow game logic over safety logic.
Level
Hype check
Tap to vote and see what everyone thinks.
