Another thought on this, On Thu, Jun 4, 2026 at 7:23 PM Jay Daley <[email protected]> wrote:
> > > > On 5 Jun 2026, at 13:15, John R Levine <[email protected]> wrote: > > > > We can certainly ask, but it's up to the contributor how much to answer. > > > > Honestly, if it turns out that we're publishing big chunks of AI slop > unedited, > > > This is based on the assumption that AI produces slop. In my experience > the more effort you put into it, the better the quality of output. It will > not be too long before many, many people can instruct AI well enough to > produce high quality output. > > Jay — Obviously my earlier theoretical premise is based on IETF publication streams (and other streams might struggle with this more with less external input before publication assessment)... But AI assisted tooling from spell check, to grammar checks, to completely reflowing and rewriting an email entirely is baked into many mainstream email clients. In fact, I would say a some 5-10% of the time I spend reviewing an email is to double check that a grammar fix has still maintained my voice before I hit send. When a large proportion of IETF work is done through email participation across various WG fora, and those contributions are heavily polished by AI already and those contributions make it into an active WG I-D…? There’s just not really a way of knowing what part of that is AI synthesis, and I think it’s sensible to carefully consider Martin Thompson’s earlier argument. That is, if I understand Martin’s point correctly, it would be to affirm any contributions to the IETF as directly associated to the named (human) author behind the email address and as such accountability for such content is applied in the same way; regardless of external tool assists may have been used in its creation. Cheers, Nathanael P.S. I didn’t apply any AI tools to help polish this email, (maybe that’s obvious or not).
-- rswg mailing list -- [email protected] To unsubscribe send an email to [email protected]
