Last week a request for comments (RFC) was issued around establishing an LLVM AI Tool Use Policy. The proposed policy would allow AI-assisted contributions to be made to this open-source compiler codebase but that there would need to be a “human in the loop” and the contributor versed enough to be able to answer questions during code review. Separately, yesterday a proposal was sent out for creating an AI-assisted fixer bot to help with Bazel build system breakage.
Last week’s LLVM AI tool policy was brought up for discussion. AI-assisted contributions would be welcome as long as there is a human in the loop that understands the code and competent enough for answering any questions during the code review. Contributors should also be transparent if there are “substantial amounts” of tool-generated content. This pull request in turn is open on GitHub for adding their AI contribution policy to the LLVM documentation. That LLVM Ai tool policy remains under discussion.
Separately, raised yesterday for request for comments was a proposal by Google compiler engineer Pranav Kant for creating an AI-assisted Bazel fixer bot. This AI bot would automatically create AI-generated pull requests for fixing broken LLVM builds when using the Bazel build system. Google is a big user of the Bazel build system but LLVM contributors are not required to update their changes around Bazel builds. Thus Google compiler engineers would like an experimental AI-assisted bot for helping to maintain their Bazel build support in LLVM.
The original proposal there calls for this AI bot to create the pull requests with fixes, but some have taken some objection to that part. Some would prefer a human to at least evaluate the proposed changes first rather than the bot creating the PRs straight-away and creating a bigger burden on LLVM developers to review that code straight from a bot.
Meanwhile GNU toolchain developers have also been considering an AI/LLM policy for the GCC compiler too.
