Torvalds vs. Debian: Two Opposite Visions for AI in Open Source

The result, merged in Linux 7.0 in April 2026, is the file Documentation/process/coding-assistants.rst — a brief, practical document that establishes three rules:

• Only humans can sign off. "AI agents MUST NOT add Signed-off-by tags. Only humans can legally certify the Developer Certificate of Origin (DCO)" . The DCO is a legally binding attestation of authorship and licensing; no AI can make it.
• Attribution via "Assisted-by" tag (recommended, not enforced). Contributors who use AI tools are asked to include an Assisted-by: tag identifying the model name and version, for example

  Assisted-by: ChatGPT-4 (gpt-4-0314) [GitHub Copilot]

  .
• Full human liability. The submitting human takes complete legal and reputational responsibility for every line, every bug, and every security flaw in AI-assisted patches .

Torvalds's July 15, 2026 post to the kernel mailing list made his underlying logic clear: AI tools are optional productivity aids, akin to compilers or debuggers, and the kernel should not become "one of those anti-AI projects" that bans useful tools . He warned of practical downsides but said maintainers should adapt tooling and review practices rather than impose top-level bans .

The Debian Project: A Formal Push to Ban LLM-Generated Contributions

Debian's approach is diametrically opposite. On July 24, 2026, the project opened the discussion period for General Resolution vote_002, "LLM usage in Debian," proposed by developer Matthias Geiger and seconded by seven contributors including Ian Jackson . The official Debian vote page states the resolution's aim to "expressly forbid any contributions to Debian written with the use or assistance of large language models (LLMs) or other generative AI tools" .

The proposed ban covers a broad scope of direct Debian contributions:

• Debian source packages
• Official Debian project software (such as lintian)
• Official documentation and translations
• Official communications of the project
• Bug reports
• Wiki content

Importantly, the resolution explicitly excludes upstream projects using LLMs for their own development, AI-related software packaged by Debian, and upstream patches or security fixes .

The General Resolution presents voters with multiple ballot options:

• Proposal A (the ban): Forbid all LLM-assisted contributions to Debian.
• Proposal B (permissive with conditions): Allow AI-assisted contributions provided certain conditions (disclosure, tagging, etc.) are met.
• Further options: Additional middle-ground or "further discussion" choices .

The resolution cites four categories of concern: legal uncertainty over the copyright status of LLM-generated output, quality and accuracy degradation risks, community health impacts including the strain on reviewers, and ethical issues tied to LLM training practices . Debian sees AI-generated code as a potential supply chain and legal risk to the distribution .

This is not Debian's first attempt at an AI policy. In February-March 2026, a draft General Resolution proposed by ex-Debian Project Leader Lucas Nussbaum would have allowed AI-assisted contributions under seven conditions including mandatory disclosure and machine-readable tags . After weeks of intense mailing list debate, the community abandoned that vote entirely, opting to continue handling AI contributions case-by-case . The July 2026 effort is a new, more restrictive proposal that some observers see as a reaction to the earlier failure to reach consensus .

Side-by-Side: The Key Differences

The Growing Rift in Open Source

These two positions represent the poles of a deepening ecosystem divide. Torvalds, as a single benevolent dictator, can make a snap call to welcome AI tools and tell dissenters to fork . Debian, as a community-governed project with strong commitments to free software principles and legal caution, is working through a formal democratic process to ban what it sees as an existential threat to human-authored open source .

Observers note this rift tracks broader open-source fragmentation: Ubuntu has embraced AI, while other projects like Curl have banned AI-generated contributions over quality concerns . The Debian vote, should it pass, could set a powerful precedent for how major Linux distributions handle AI-generated contributions going forward .

For contributors navigating this landscape, the practical implication is clear: what is welcomed in the Linux kernel could be a policy violation in Debian. The open source ecosystem no longer has a single position on AI — it has a growing number of competing ones, and each project's choice reflects its deepest governance values.