Re: Ballot option: Allow AI-Assisted Contributions
"Theodore Tso" <[email protected]>
| Newsgroups | gmane.linux.debian.devel.vote |
|---|---|
| Message-ID | <[email protected]> |
On Thu, Aug 06, 2026 at 11:02:53PM -0500, Simon Richter wrote: > Now, there are multiple possible interpretations: Sorry, none of this makes any sense. > 1. the output is not a derivative work of the context. The prompter > therefore is not the author, and asserting that they own the copyright is > therefore fraudulent. In the case where the person is submitting a changes to most projects, they don't need assert that they own the copyright. This might be true in the case of (say) the FSF copyright assignment, but I refuse to sign the FSF copyright assignment document, mainly because I know what the word "indemnify" means, and I own a house. For most open source projects, you just have to assert that you have the right to submit the change, not that you own the copyright. This is critical when you have multiple contributors, potentially working for different employers, and so asserting that the patch submitter has the copyright would be highly problematic. Take a look at the Developers Certification of Origin (which *was* drafted by lawyers, not random people making opinions about how copyright law works :-). > 2. the output is a derivative work of the context. We therefore need a > licence from the service provider to be able to freely choose a licence for > the generated output (I suppose we have that, but we need to make sure), and > the generated code is largely attributable to the authors of the existing > code, not the person implementing the latest change. Again, so what? Someone might clean up existing code base by refactoring the code, and/or moving chunks of code around --- say to avoid a use after free bug. The fact that the patch is largely attributable to the authors of the existing code, not the person implemented the latest change is a feature, not a bug. And again, most open source projects don't forbid that, because we like code cleanups and paying down technical debt. This isn't a problem without AI. Adding AI doesn't magically make things problematic. If people were to hold human beings to the same standard as what people are trying to assert for LLM's, open source would be dead in the water. Especially with the existence of Stack Exchange.... - Ted