On Thu, Jun 4, 2026 at 1:59 PM John R Levine <[email protected]> wrote:
> > For me the issue is an ethical one. It feels wrong not to disclose if > > an AI wrote my homework. I think we need to put that point to the > community. > > If the community doesn't want such a rule, fine, but IMHO it needs to be > a > > wider community than the RSWG that reaches that conclusion.. > > We have always allowed people to use material they didn't write in > contributions and we have never required people to say where the material > came from. What's different now? If people have to disclose LLM > material, do they have to disclose other PD material they use? There has > always been a possibility that people will assume that an author wrote > some chunk of a document that he used from somwhere else. > This roughly matches my view. I think it's important to distinguish two quite different purposes for this kind of notice: * Giving credit to *others* where it's due. * Disclaiming that the work is not entirely one's own. As a general matter, IETF specifications are conventionally the product of the WG, not solely of any individual author, regardless of whose name appears on the front, so I don't really see how an AI disclosure is necessary to serve the second purpose. As for the first, AI isn't -- at least at present -- a person who needs to be given credit, but in any case It's quite common to have extensive material contributed by unnamed WG members. For example, consider RFC 2821's Acknowledgements section: Many people worked long and hard on the many iterations of this document. There was wide-ranging debate in the IETF DRUMS Working Group, both on its mailing list and in face to face discussions, about many technical issues and the role of a revised standard for Internet mail transport, and many contributors helped form the wording in this specification. The hundreds of participants in the many discussions since RFC 821 <https://www.rfc-editor.org/info/rfc821/> was produced are too numerous to mention, but they all helped this document become what it is. I think this reflects the understanding that most RFCs are WG efforts. This isn't to say that I don't think acknowledgements are good (hence the extensive Contributors list in TLS 1.3) but merely that I don't think it's somehow misleading not to have them, let alone not to list that you had AI assistance. In addition, Brian writes: > If the community doesn't want such a rule, fine, but IMHO it needs to be a > wider community than the RSWG that reaches that conclusion.. This is backwards. No such rule currently exists and so any change to require one would need the broader community to come to consensus that one is needed, *not* the community to come to the conclusion that one is not needed. -Ekr
-- rswg mailing list -- [email protected] To unsubscribe send an email to [email protected]
