On Tue, May 26, 2026 at 2:30 PM Brian E Carpenter < [email protected]> wrote:
> On 27-May-26 05:41, Eric Rescorla wrote: > > > > > > On Tue, May 26, 2026 at 12:32 AM Brian E Carpenter < > [email protected] <mailto:[email protected]>> wrote: > > > > > > > Section 8 talks about tools, but the responsibility razor cleanly > applies to the text about plagiarism. If content has been plagiarized, the > authors are responsible. The authors might then attempt to hold a > contributor responsible, but there is limit to how much responsibility can > be deflected in these cases. > > > > Yes, which is exactly why it says "The authors or editors remain > entirely responsible for any content generated by AI" and "The authors or > editors must verify that no unacceptable plagiarism has been performed by > AI." You're right, the same applies to any form of contribution, not just > AI contributions. > > > > > > How, precisely, do you expect that authors are to do that? > > Well, as an active author for an IEEE publication, I have recently > submitted an article so I can tell you my procedure - I asked my coauthors > if they'd used AI. > > > Suppose that Alice is an editor of an RFC and Bob provides a PR > with some new text, which turns out to have been plagiarized from some > non-RFC publication by Charlie. What is it you believe Alice ought to have > done in order to detect this? > > Well, there are tools for detecting plagiarism (very necessary for anyone > evaluating student assignments these days) and AI-written text. I'm not even talking about AI-written text here. Just simple plagiariam. > But it comes back to Martin's insistence on author/editor responsibility - > that's Alice's problem, and shouldn't become the RPC's problem. > That's a non-answer. I've submitted many PRs to I-Ds and I can't think of any instance in which anyone has asked me if my text was original, and I doubt very much they are running plagiarism detectors, so to the extent to which this is the author's responsibility, I don't see much evidence they are fulfilling it. If you are creating a new requirement, I think you should first demonstrate it's practical. This is also another reason we'll need review by counsel. I don't think that we need counsel to tell us whether this idea is practical. -Ekr
-- rswg mailing list -- [email protected] To unsubscribe send an email to [email protected]
