proposed AI policy

[Bryn M. Reeves]

On Wed, Apr 08, 2026 at 05:49:09PM -0400, Aaron Merey via Systemtap wrote:
> I support this policy. The existing patch review practices are capable
> of handling AI-assisted contributions much the same as contributions
> that are not AI-assisted. Low quality submissions can be refused,
> commits found to have copyright or license problems can be reverted,
> etc.

Agreed. We have something similar in the projects I maintain (which has
been picked up by a couple of other projects since):

  https://github.com/snapshotmanager/snapm/blob/main/AI_POLICY.md

> Should the policy include a mandate that contributors disclose when a
> contribution is AI-assisted? In this case I'm leaning towards no
> mandatory disclosure. The stated rationale for this policy is that
> contributions speak for themselves regardless of tool use or creator
> identity. So why should we demand special disclosure rules for just
> one specific type of tool? There is no settled precedent for mandatory
> disclosure so AFAIK we don't have a legal requirement to demand this
> from contributors. Like any other contributor policy we can update
> requirements if needed.

We ask that contributors disclose ("If you do use AI assistants to write
meaningful volumes of code, we prefer that you make this clear...") but
we stop short of mandating it.

If users do decide to mention it we ask that they use a tag like this
in their commits for consistency:

  Assisted-by: Some AI Tool https://someai.example.com

We have a few dozen such commits in the tree right now from a variety of
authors.

Best,
bmr.