rust-lang/rust is adopting an LLM policy

Published: August 5, 2026 — Inside Rust Blog (Jynn Nelson)

Five teams in the Rust project (compiler, libs, types, rustdoc, and bootstrap) adopted a formal policy governing how Large Language Models can be used when contributing to the rust-lang/rust monorepo. The policy is not an official project-wide stance on LLMs and does not apply everywhere in the Rust project.

Core principle

It’s fine to use LLMs to answer questions, analyze, distill, refine, check, suggest, review. But not to create.

Why

LLM-assisted contributions had created issues: polished PRs no longer signal effort/understanding; easier code writing worsens review bandwidth (~1,281 open PRs cited); mechanical copy-paste of LLM replies wastes reviewer time. Prior moderation was unpublished “wild west”; the new policy makes rules public and consistent.

Key rules

• Private use without disclosure allowed if content is not posted for others to read/review
• Banned: LLM-created comments, docs, or diagnostics; replacing human judgement; requiring LLMs to contribute
• Disclosure required for machine translation, trivial changes, bug discovery, and LLM-assisted reviews when posted
• LLM-created code: conditionally allowed under experiment — pre-arranged with a named reviewer, non-critical to soundness, well-tested, well-reviewed, fully disclosed; ai-assisted label
• Circuit breaker: if more than half of PRs merged in a 6-week window are LLM-created, merging new LLM-created PRs pauses until below 50% (minimum 10-day cooldown)
• Reviewers may close non-compliant PRs; no one is required to review LLM PRs unless they volunteer
• Harassment for LLM use is prohibited under the Code of Conduct

Scope limits

Does not cover other rust-lang org repos, language-team work (tracking issues, stabilization), style guide, or non-ratifying teams. Project-wide policy still discussed as desirable future work.