Debian has adopted a policy that neither endorses nor prohibits generative AI in the development, maintenance or documentation of its Linux distribution. Voting developers rejected alternative proposals that would have barred AI-assisted contributions.
Instead of creating a separate standard for generated work, Debian will hold every submission to its existing requirements for correctness, quality, maintainability and legal compliance. The person submitting a change remains responsible for understanding, reviewing, testing and, when necessary, modifying any AI-assisted output. Uploading generated material without meaningful human review is inconsistent with the project’s development practices.
Contributors are encouraged to disclose when they used AI, but disclosure is not mandatory. That choice has drawn opposition from some community members; one contributor said he was leaving the project. Ubuntu developer Canonical faced similar criticism earlier this year over its own approach.
The policy gives maintainers flexibility to use coding assistants while avoiding a blanket exception for their mistakes. It also leaves reviewers without a guaranteed label showing which changes received machine assistance. In practice, Debian is treating AI like another development tool: acceptable when the contributor can defend the result, but not a reason to relax licensing, testing or maintenance obligations.