OpenAI says its AI found 948 Linux flaws and wrote an exploit for every one

OpenAI said its AI system identified 948 vulnerabilities in the Linux kernel and produced a working exploit for each one. The finding points to a widening role for AI in software security, where automated discovery and exploitation can outpace manual review and traditional patching cycles. The result matters because Linux underpins much of the server, cloud and crypto infrastructure that digital-asset markets depend on, so faster flaw detection could compress the window between disclosure and attack. What to watch next is how quickly maintainers triage and patch the reported flaws, and whether AI-driven auditing becomes a standard layer in security workflows across major open-source projects.