Linux Kernel Code Removals Driven by LLM-Generated Security Reports

Summary

LWN.net reports on a growing problem within the Linux kernel development community where LLM-generated security vulnerability reports are driving the removal of code, primarily from the networking subsystem. The reports, apparently produced by large language models, are flagging code as vulnerable and prompting maintainers to remove or refactor kernel components, raising serious concerns about the reliability and trustworthiness of AI-generated security assessments in critical open-source infras