rust-lang/rust introduces LLM usage policy

Five teams within the Rust project have adopted a policy governing the use of Large Language Models (LLMs) when contributing to the rust-lang/rust monorepo. This policy does not represent an official stance for the entire Rust project and currently applies only to rust-lang/rust.

The policy applies to the following individuals:

• Those who review or mediate PRs in rust-lang/rust
• Those who author PRs, including those containing LLM-generated code
• Those who use LLMs to discover and file issues
• Those who write issues or comments by directly quoting LLM responses

The introduction of this policy stems from concerns that the increasing use of LLMs is disrupting traditional open-source collaboration methods. Specifically, three major issues have been highlighted:

• High-quality outputs no longer reflect actual effort and understanding
• The ease of writing code has increased the burden on reviewers and the backlog of PRs awaiting review
• Mechanically copying and pasting between LLMs and repositories wastes community time

At the time the policy was drafted, the Rust project had 1,281 open PRs, and it was noted that the ease of LLM-based code generation has widened the gap with already limited review capacity.

The new policy aims to formalize previously informal dedicated channels and mediation criteria into public rules, thereby reducing instances where new contributors have their PRs closed without explanation and enabling reviewers to respond based on established policy.