Linux kernel maintainers reportedly decided to deprecate the Assisted-by tag, which attributed AI-tool contributions to patches. The Register reported on October 23, 2026, that Sasha Levin announced the decision at the Linux Kernel Maintainers Summit in Prague. In comments reported by the same outlet, Linus Torvalds said competent developers should be able to use AI as a tool, while arguing against patches from contributors who do not understand the code or intend to take part in the project.
Torvalds ties AI use to understanding the code
Torvalds’s position favors developers who use AI while learning and contributing to the kernel. He distinguished them from people who submit a patch without understanding its code or intending to become involved in the community. His support for AI assistance does not amount to automatic approval of AI-generated patches.
Torvalds also described two pressures: AI can find bugs in the kernel’s large codebase, adding work for maintainers, and contributors can use AI to generate patches they do not understand.
What the Assisted-by decision concerns
The reported decision concerns an attribution line identifying AI-tool contributions. It is separate from Torvalds’s comments about whether contributors understand their code and intend to participate in the project.
What earlier kernel guidance covered
On December 23, 2025, a Linux kernel commit added documentation for AI coding assistants covering licensing, Signed-off-by requirements, and attribution format. The official commit record provides that earlier policy context.
Kroah-Hartman’s view of AI-generated security reports
Linux kernel maintainer Greg Kroah-Hartman described a shift in the security reports he had seen, from low-quality AI-generated reports to reports he considered real and useful. That is his assessment of the reports, not a measured project-wide result.
