AI-Generated Security Reports Are Now Deleting Code From Linux Kernel

Here's something wild: security researchers are using large language models (LLMs) to scan the Linux kernel for vulnerabilities, and when they find issues, those AI-flagged problems are actually getting code removed from the kernel itself. It's like having an AI security guard that doesn't just point out problems—it convinces the maintainers to delete the suspicious code entirely. The thing is, we're not totally sure how accurate these AI reports are, or whether they're catching real threats or just creating false alarms that lead to unnecessary deletions. This raises a pretty big question: should we really be letting AI systems drive decisions about what code stays and what gets cut from one of the world's most critical software projects?