Debian Puts LLM Usage Policy to a Vote: Three Proposals
Original: LLM Usage in Debian: Three Proposals
Why This Matters
Debian's LLM policy decision could set a precedent for how major open-source distributions govern AI-generated contributions.
The Debian Project has opened a General Resolution vote on the use of Large Language Models (LLMs) within the project. Three competing proposals are on the ballot, reflecting differing views among Debian developers on how AI-generated content should be handled in the distribution.
The Debian Project has initiated a formal General Resolution (GR) vote titled 'LLM usage in Debian,' presenting members with three distinct proposals to determine official policy on the use of Large Language Models. The vote page is listed under the 2026 voting cycle and is currently in the discussion or active voting phase. The three proposals (designated Proposal A and at least two others) represent varying stances — likely ranging from permissive to restrictive — on whether and how LLM-generated content may be incorporated into Debian packages, documentation, or project contributions. This GR follows a previously withdrawn resolution titled 'Interpretation of DFSG on Artificial Intelligence (AI) Models,' suggesting the project has been actively wrestling with AI policy for some time. Debian uses the Condorcet/Clone Proof SSD voting method for such resolutions, requiring a supermajority or simple majority depending on the proposal type. No specific vote counts or outcome have been published yet, as the resolution appears to still be in progress.