Open AI Model Turned Firefox Fixes Into Working Attacks
Research group Epoch AI reports that GLM 5.3, a widely available AI model, turned public Firefox security fixes into working attacks. It compromised six Firefox releases in under 48 hours, at roughly $1,000 per release, with no human input and without tripping safeguards. The finding suggests users may be exposed before updates reach them.
Quick glossaryEpoch AI, open-weight model, exploit chains, API fees
- Epoch AI
- A US-based research group that studies and tracks trends in artificial intelligence.
- open-weight model
- An AI model whose trained settings are published so anyone can download and run it.
- exploit chains
- Linked steps that take advantage of software flaws in order to break into a computer system.
- API fees
- Charges paid to use an AI company's software gateway that lets other programs send it requests.
Burning Potato score: 81%