LLVM Considers Banning AI-Generated Pull Request Descriptions and Text Communication

Reported 5 Oct 2026 · entered the ledger 19 Oct 2026 · last checked 19 Oct 2026 · Side effect

LLVM maintainers faced an influx of verbose, low-quality AI-generated pull request descriptions and RFCs that drained limited reviewer bandwidth. Because prior policy relying on human-in-the-loop responsibility and encouraging contributors to write their own descriptions proved ineffective, maintainers proposed a formal ban on AI-generated text communication.

The institution LLVM Project institution

The mechanism Friction removed externality scale 2 of 5, fast

Adaptation reform the rule itself was rewritten

The case as the ledger holds it

The assumption that broke: Writing pull request descriptions was assumed to require effort and a genuine understanding of the code, acting as a natural filter and sanity check for incoming contributions.
The first-order effect: A high volume of verbose, low-quality AI-generated PR descriptions and RFCs entered the review queue, requiring disproportionate maintainer time to read and manually edit into mergeable commit messages.
Who pays: LLVM code reviewers and open-source maintainers whose review bandwidth is consumed reading and editing generated text · group: Volunteers and reviewers
Scale and speed: 2 of 5 · fast

Evidence 1 page read

LLVM Discussion Forums ↗ 5 Oct 2026
Such pull request description are a drain on our limited reviewer bandwidth, both because they take more time to read in the first place, but also require substantial massaging to bring them into a mergeable form.

The second reader every claim, checked against the page

Claim	Verdict	Note
[date] The RFC was published on October 5, 2026.	supported	Page metadata gives Published Time: 2026-10-05T14:32:44+00:00.
[fact] LLVM's existing AI tool policy allows essentially all uses of AI as long as there is a 'human in the loop' who takes responsibility.	supported	Page: 'LLVM’s AI tool policy is fairly liberal and allows essentially all uses of AI, as long as there is a “human in the loop”, who takes responsibility for the contribution.'
[fact] The project receives a 'very large amount of low-quality pull request descriptions', especially from new contributors, that are verbose and contain unnecessary information.	supported	Page: 'the very large amount of low-quality pull request descriptions we receive, especially from new contributors. These are often extremely verbose, contain huge amounts of unnecessary information...'
[quote] AI-generated pull request descriptions 'are a drain on our limited reviewer bandwidth, both because they take more time to read in the first place, but also require substantial massaging to bring them into a mergeable form.'	supported	Page: 'Such pull request description are a drain on our limited reviewer bandwidth, both because they take more time to read in the first place, but also require substantial massaging to bring them into a mergeable form.'
[quote] The previous policy guidance that 'strongly encourages' writing PR descriptions oneself 'turned out to be essentially useless.'	supported	Page: 'our current AI tool policy “strongly encourages” writing PR descriptions yourself turned out to be essentially useless.'
[fact] The RFC proposes to ban the use of AI-generated or substantially AI-assisted textual communication, including PR descriptions and RFC texts, while leaving allowable AI uses for code contributions unchanged.	supported	Page: 'proposes to ban the use of AI-generated or substantially AI-assisted textual communication, including, but not limited to, for pull request descriptions and RFC texts. It does not make any changes to the allowable uses of AI for code contributions.'