Debian dismisses AI-contributions policy

Posted May 13, 2024 2:33 UTC (Mon) by raven667 (subscriber, #5198)
In reply to: Debian dismisses AI-contributions policy by NYKevin
Parent article: Debian dismisses AI-contributions policy

> You also need to define what exactly counts as an "AI contribution."

You can define a little but there is more flexibility if the rules are open to interpretation, state some principles, don't proactively police it, and when something becomes a problem you have a rule to point back to, otherwise if the rules are over-specified clever people will try to craft procedural "workarounds" to ignore the spirit of the rule, if the rules are flexible then it's really up to consensus opinion whether an enforcement action is reasonable or not, rather than arguing about minutia, making a case to the community on the substance of the issue, if they dont agree they'll let you know. You don't have to proactively police with perfect detection to be able to enforce rules, they can applied on as-needed basis and rely on the judgement of the people involved. Written rules can't replace judgement anyway.

Debian dismisses AI-contributions policy

Posted May 13, 2024 8:25 UTC (Mon) by NYKevin (subscriber, #129325) [Link] (2 responses)

Yes, but you still need a *generalized* agreement on the overall category of conduct you want to prohibit. There is a lot of distance between "writing some documentation in broken English, asking ChatGPT or Grammarly to fix the grammatical errors, and submitting the original broken English alongside the AI-improved one so that maintainers can verify that no misinformation was introduced," and "having Copilot write a whole function from scratch and firing off a YOLO patch to the mailing list (or forge or whatever)." If you just say "AI contributions are banned, no we will not tell you what that means," then that will not result in civilized discourse. It will result in people disagreeing on what the policy means, probably at the top of their lungs, and regularly accusing one another of exactly the kind of chicanery you say you want to avoid.

I'm not saying there needs to be some perfectly decidable algorithm for determining what does or does not count. I'm just saying, you need general principles which are more specific than "AI bad, do not use."

Debian dismisses AI-contributions policy

Posted May 13, 2024 18:10 UTC (Mon) by rgmoore (✭ supporter ✭, #75) [Link] (1 responses)

It seems to me that the basic policy should be that users are responsible for everything they submit. That means two big things:

The submitter is responsible for the quality of their submission. If they are submitting junk, they can expect to be treated as an annoyance. That's true regardless of why it's junk. It doesn't matter if they're a bad programmer, they're depending on an incompetent AI, or whatever. Send in enough junk, and you can expect your submissions to be forwarded to the bit bucket unread.
The submitter is responsible for ensuring they have the right to submit their contribution in the first place. This means they must either hold copyright themselves, have a valid license from someone else that gives them the right to submit it, or know it isn't covered by copyright. Right now, the copyright status of AI output needs clarification, which means nobody can be sure they have the right to submit AI output.

Debian dismisses AI-contributions policy

Posted May 13, 2024 18:27 UTC (Mon) by mb (subscriber, #50428) [Link]

> users are responsible for everything they submit.

2 is basically that:

https://en.wikipedia.org/wiki/Developer_Certificate_of_Or...

And 1 is, well. Common sense and common practice.

>Right now, the copyright status of AI output needs clarification

For some people here the state "obvious" ;-)