AI Preferences
charter-ietf-aipref-01

[Ballot block]
I am confused about the combination of:

    such as the Robots Exclusion Protocol (RFC 9309)

With:

    [out of scope]
    Authenticating or authorising clients and/or crawlers

robots.txt is explicitly about authorization. So this is now both in scope and out of scope ?

[Ballot block]
An easy to fix BLOCK ballot: let's be specific about the intended status in the charter, which has IETF community consensus, rather than in datatracker milestones, which has no IETF community consensus.

Changed charter milestone "A standard track protocol specification to associate the expressed AI-related preferences to the IESG for publication", set description to "Standard track protocol specification(s) to associate the AI-related preferences to the IESG for publication", set due date to August 2025 from December 2025

[Ballot comment]
“A means of attaching or associating those preferences to content in IETF-defined protocols and formats, including but not limited to using Well-Known URIs (RFC 8615), Robots Exclusion Protocol (RFC 9309), and HTTP response header fields."
A method for reconciling multiple expressions of preferences.”

There are editorial or formatting issues. There is an extra quote mark after header field. The “A method for reconciling …” sentence doesn’t flow.

“A standard track protocol specification to associate the expressed AI-related preferences to the IESG for publication”

Given the vagueness of the text does this milestone imply that exactly one PS protocol specification will be produced?

What protocols or formats will it cover?

[Ballot comment]
Just some minor comments (hence preventing a strong YES):

Even if "AI" is a well-known acronym, suggest to use "Artificial Intelligence" at least in the charter text (and possibly in the WG name).

`AI model development, deployment, and use`, while I understand the "development" part (suggest using "training" ?), I wonder what is meant here about "deployment, use".

`associating those preferences to content in IETF-defined protocols` it took me a while to parse the sentence correctly as I first read it as "content of the IETF-defined protocols". It may be my French bias that cause this misread of mine.

`header fields."` suggest removing the extra "

`IPTC and PLUS` as I am probably not the only one not aware of these groups/SDO, suggest expanding the acronyms.

`Liaisons are intended to inform external bodies ` seems to indicate a one-way communication from IETF to other bodies and not listening to their feedback/requirements

This came out of IAB workshop on AI-Control, later it was discussed in a side-metting at IETF121. Given the discussion it already had and given the interests at the side-meeting to work on this at IETF, the decision was to create a working group in IETF.