Kernel Patches Hide AI Usage

The Linux kernel project has removed the requirement that contributors tag code assisted by large language models, so maintainers can submit LLM‑generated or LLM‑refined patches without declaring that assistance. That breaks an emerging norm of provenance for code authorship and makes it harder for reviewers, downstream distributors, and security auditors to trace when and how AI influenced kernel logic. — This matters because the kernel powers billions of devices; burying AI provenance raises risks for code quality, supply‑chain trust, vulnerability attribution, and governance of critical infrastructure.