Re: GCC GSoC 2026: Call for project ideas and mentors
Richard Biener <[email protected]>
| Newsgroups | gmane.comp.gcc.rust,gmane.comp.gcc.devel,gmane.comp.gcc.fortran,gmane.comp.gcc.libstdc++.devel |
|---|---|
| Message-ID | <CAFiYyc3gi2v+1g3eVYt9QuEzD62ym6S_wDZsKd+Vcdm4gmgLpg@mail.gmail.com> |
On Mon, Jan 19, 2026 at 7:46 PM Ville Voutilainen <[email protected]> wrote: > > On Mon, 19 Jan 2026 at 20:36, Richard Biener <[email protected]> wrote: > > If GSoC does not have, we should, at least, set a clear expectation as to > > how usage of AI in completing the project is [not] allowed and should be > > documented. > > Yes, please. The GSoC mentors' discussions I can find in my inbox > don't exactly suggest such a clear policy > being in place, and various participating organizations have their > own. And I would venture a suggestion > that such a policy should be there in general for GCC, not just for > GCC GSoC projects. The DCO, even though I think it's obvious, might also get a clarifying amendment that what an AI tool produced is not obviously your work or can be considered OK to contribute. Same holds for people contributing using FSF assignment - you can only assign sth you have rights on. Richard. > In various places, people are being reminded, rather than newly told, > that blindly merging LLM-generated code > is as bad (and in some ways exactly as bad) as blindly copying code > from a random internet site, even if the > name of that site is "StackOverflow" and you really really like the > site. The problems are identical, and such blind copying can't be > accepted, > due to copyright reasons.