It'll come back to personal relations more (was: Re: Controlling the flood of 'AI slop' independent Internet-Draft submissions)

[Lars Eggert <[e-mail]>]

Hi,

I believe the issue of fully or partially AI-generated IETF contributions is here to stay.

(Let’s not open the can of worms of whether they *can* actually be IETF contributions, given that we require licenses be granted to the IPMC.)

I don’t see  a future in which the AI genie is fully being put back into the bottle, and AI can be useful in accelerating standards work if used responsibly and correctly.

(Again, without opening the cans of worms on whether how training material was/is being sourced, what AI is doing to society in general, etc., etc.)

I also don't think that tooling, -00 submissions restrictions, etc. can be more than a stopgap. I think they will help some with preventing DDoS-like events. But I also think we’re just generally going to see more contributions by more participants - which is not inherently bad, we’ve been discussing how to grow the community for decades now.

The issue is of course the initial quality of those new contributions, and the amount of community time needed to review and vet them. This is pretty much the exact problem we have in open source, where PRs are now cheap to generate and the cost is shifting to core review, which is not (yet?) similarly cheaply done.

For myself, this will probably mean that I’ll be quicker to ignore or disregard new contributions from folks I have not had a history with, unless maybe I see someone else I do have a history with meaningfully engage. Or whether the contributions is squarely in an area I care a lot about.

If more people in the community will do this, which I think is not unreasonable, the downside may be that the IETF will become *less* open to new participants, and not more. Genuine new participants with worthwhile ideas will have a harder time being spotted in the “AI slop”, and that’d be a shame. How can we fix *that*?

Thanks,
Lars