Unattributed: Exploit

An automated OpenAI system produced working exploits and draft fixes at roughly $1,279 per flaw, raising new questions about how cheaply AI can find holes in critical software at scale.

Evidence boundary

Confidence describes the coverage of the available evidence. It is not a safety rating and does not guarantee that a protocol or asset is safe.

Sources

• https://decrypt.co/382632/openai-linux-security-flaws-exploit

Record history

First seen: 2026-11-05T19:16:03.000Z. Last updated: 2026-11-06T20:32:40.176Z. Revision: 4.