Debian Project adopts policy for responsible generative AI use

The Debian Project has adopted a new policy regarding the use of generative artificial intelligence, opting for a middle-ground approach titled ‘Responsible Use of Generative AI’. Following a two-week voting period involving nine different proposals, project members selected this option to balance productivity gains with quality control.

The policy does not explicitly mandate or prohibit the use of generative AI tools for the development, maintenance, or documentation of software, packages, and other media within the project. While acknowledging that these tools can significantly increase contributor productivity, the project maintains that all submissions must meet the same rigorous standards for quality, correctness, maintainability, and legal compliance, regardless of the tools used to create them.

During the Condorcet voting process, the chosen proposal defeated several other options, including a total ban on LLM-assisted contributions via the Debian Social Contract and various restrictive conditional models.