RE: Dealing with LLMs in IETF discussions draft

[Andrew Campling <[e-mail]>]

The RASPRG presentations were helpful in highlighting several challenges that are likely relevant across many, if not all, SDOs. Lowering barriers to participation through AI can be beneficial, particularly if it broadens and diversifies the contributor base. However, these same tools can also be misused, making it easier to overwhelm a mailing list, working group, or even an entire SDO with what effectively amounts to a denial-of-service-style attack.

As I noted during the RASPRG session, this raises the question of whether our rules of participation need updating. They were likely developed under the assumption that all participants would engage in good faith, an assumption that may no longer apply, for example, if some "participants" are actually AI tools auto-responding to mailing lists.


Andrew

From: Ignacio De Castro Arribas <[e-mail]>
Sent: 31 July 2026 15:58
To: Stephen Farrell <stephen.farrell=[e-mail]>; IETF-Discussion <[e-mail]>
Subject: Re: Dealing with LLMs in IETF discussions draft

This is definitely interesting, and it is a fact that AI use is becoming almost omnipresent.

There have been multiple discussions on the topic at RASPRG, including the last one in Vienna with two presentations on how-to/tools for participation from Mark Nottingham and inferring AI usage in drafts from Jaime Jiménez.
https://datatracker.ietf.org/meeting/126/session/rasprg
Given the amount of discussion we did a poll towards making an interim to keep discussing the topic (ie, IA-enabled participation at the IETF).

Cheers,

Ignacio



---

Dr. Ignacio Castro

Assistant Professor

Electronic Engineering and Computer Science

Queen Mary University of London



________________________________
From: Stephen Farrell
Sent: Friday, July 31, 2026 04:20
To: IETF-Discussion
Subject: Dealing with LLMs in IETF discussions draft


Hi all,

Myself and Chong Feng have posted an I-D [1] that (from two
pretty different perspectives) aims to progress discussion
of how we might better handle LLM text in IETF discussions.

I think this'd be a fine topic for broad discussion. I do
not know what would be the correct mailing list for such a
discussion, but also think it doesn't warrant waiting for
another round of dispatchery before being discussed. (Nor
would any dispatch outcome above 'new mailing list' suit.)

So, I'd be v. interested in comments on this and/or on it
being moved to some other mailing list for discussion. If
no existing list suits, I'd suggest [e-mail]<mailto:[e-mail]>
(which could be contracted to [e-mail]:-<mailto:[e-mail]:->) as a name
for a new list, but whatever works is fine.

Meanwhile, we'd be v. interested in any comments here or
on the github repo [2].

And of course, this is a v. rough thing, not intended to
be prescriptive or authoritative, so have at it:-)

Cheers,
S.

[1] https://datatracker.ietf.org/doc/draft-fengfar-led/
[2] https://github.com/sftcd/led/