Linus Torvalds has a low tolerance for rules that police tools rather than outcomes. That came through clearly in a recent Linux kernel mailing list thread about whether to keep the Assisted-by: LLM disclosure tag on AI-assisted patches.
Why the Tag Is Getting Dropped
Kernel developer Sasha Levin proposed removing the tag after data showed it was not working. Of the 18,533 non-merge commits in Linux 7.3 through release candidate 7, about 10% included an Assisted-by tag. But AI tools are believed to be in use far more widely than that. Many contributors do not know the requirement exists. Others skip it deliberately. The tag tells maintainers almost nothing reliable.
The proposal would remove the requirement from the kernel’s contribution documentation and eliminate its handling in the checkpatch.pl script. Greg Kroah-Hartman, who previously supported keeping the tag, now backs its removal.
What Torvalds Said
When developer Laurent Pinchart argued that clear AI use expectations were still needed even if contributors ignored them, Torvalds cut straight to his position.
“I don’t understand this whole idiotic concept of ‘don’t use AI for X.'”
He also suggested, sarcastically, that developers should be banned from compilers and required to write assembly to prove they understand their code. His core position: competence is what matters, not which tools a developer used to get there.
“We’re not going to make stupid rules that say ‘you can’t use tools for X.'”
What Stays in Place
Dropping the tag does not mean AI-generated code gets a free pass. Existing documentation on generated content stays unchanged. Contributors still carry full responsibility for their submissions, including compliance with the Developer Certificate of Origin. The change removes a disclosure checkbox that was not working, not the accountability underneath it.
The revised patch has support from several prominent maintainers, but it has not yet been merged. It is a proposal with momentum, not a done deal.
