AIBleepingComputerabout 2 hours ago
New BioShocking attack manipulates AI browser into data theft
1 min read
LayerX researchers devised a prompt injection attack called BioShocking that tricks AI browsers into ignoring safety guardrails. Tested against six agentic browsers including ChatGPT Atlas and Claude Chrome plugin, only OpenAI implemented a working fix. The attack uses a fictional game scenario to make agents share sensitive data.
Level
Hype check
Tap to vote and see what everyone thinks.
