Hello, This has been a fascinating discussion so far.
If a document so filled with AI-generated synthesis were successfully adopted by a working group, it would still need to proceed through the normal means of accepting feedback from participants and elicit rough consensus, right? The document would have to survive WGLC, successive AD evaluation, and onward to IETF Last Call, before finally facing IESG evaluation. Because of that process, the finished work would surely be so significantly impacted by successive community input that it would almost certainly be utterly transformed into something demonstrably and creatively shaped by human minds. Would this not render the question of AI origin a non-issue by the time of RFC publication? At least, it certainly seems to me that it would be a truly bizarre situation if a significant portion of an otherwise publishable draft survived all the way through where enough noticeable AI-generated text would remain to actually represent a significant copyright issue. Kind regards, Nathanael On Thu, Jun 4, 2026 at 4:48 PM Jay Daley <[email protected]> wrote: > > > > On 5 Jun 2026, at 10:40, Brian E Carpenter <[email protected]> > wrote: > > > > On 05-Jun-26 10:26, Jay Daley wrote: > >>> On 5 Jun 2026, at 08:59, John R Levine <[email protected]> wrote: > >>> > >>> On Fri, 5 Jun 2026, Brian E Carpenter wrote: > >>>> As I already said, I don't think we should drag the RSWG/RSAB into > copyright > >>>> issues. We have the IETF Trust/IPMC LLC to deal with that (and impose > a > >>>> disclosure requirement as part of their Legal Provisions if they > decide it's > >>>> legally necessary). > >>> > >>> Agreed, and it isn't. > >>> > >>>> 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. > >> I don’t think the "who wrote it" is the right lens for this. I suspect > that the big issue here is the likely huge increase in uncopyrightable text > in our documents. Sure we already have public domain/uncopyrighted text, > but from my limited viewpoint it seems minor and rare, whereas we can > foresee a future where it is major and commonplace. When that happens, I > don’t think our rules/processes will be adequate and an evolution will be > needed to address that. To give you a concrete example, we prevent > non-IETF derivatives of IETF standards for good reasons, but if/when a > standard has significant AI output incorporated into it, even if that’s > light editing, will we be able to continue with that level of protection? > >> With that in mind, the key disclosure required is not if an AI wrote > it, but about the copyrightable nature of the submissions. > > > > But that's really Trust/IPMC scope, surely? > > > Implementing the policy the IETF decides is their role yes, but there is a > policy gap here because of the strategic consequences of this evolution, > and policy is for the community to decide. > > > And in the end, our only defence is going to be the RFC "brand". The law > won't help. > > I think it’s too early to tell where/how this will develop. > > Jay > > > > > Brian > > > -- > Jay Daley > IETF Executive Director > [email protected] > > -- > rswg mailing list -- [email protected] > To unsubscribe send an email to [email protected] >
-- rswg mailing list -- [email protected] To unsubscribe send an email to [email protected]
