[ai-control] Re: -04 drafts for evaluation

[Leonard Rosenthol <[e-mail]>]

>we wouldn’t be expected to develop vocabulary that operates or must be implemented at the asset level, correct?
>
Actually, yes, the group is indeed chartered to do that.

To quote from https://datatracker.ietf.org/doc/charter-ietf-aipref/,<https://datatracker.ietf.org/doc/charter-ietf-aipref/%E2%80%8B> with bold facing mine (to highlight key points).


The Working Group will focus on attaching preferences to content either by including preferences in content metadata or by signaling preferences using the protocol that delivers content.


The Working Group will deliver:

     *   A standard track document covering vocabulary for expressing AI-related preferences, independent of how those preferences are associated with content.
     *   Standard track document(s) describing means of attaching or associating those preferences with content in IETF-defined protocols and formats, including but not limited to using Well-Known URIs (RFC 8615<https://datatracker.ietf.org/doc/rfc8615/>) such as the Robots Exclusion Protocol (RFC 9309<https://datatracker.ietf.org/doc/rfc9309/>) and HTTP response header fields.

Leonard
From: Farzaneh Badiei <[e-mail]>
Date: Tuesday, November 4, 2025 at 2:27 PM
To: Lila Bailey <[e-mail]>
Cc: Sebastian Posth <[e-mail]>, Leonard Rosenthol <[e-mail]>, [e-mail] <[e-mail]>
Subject: Re: [ai-control] Re: -04 drafts for evaluation


EXTERNAL: Use caution when clicking on links or opening attachments.


Yes, thank you.
I’d like to ask a quick clarifying question. To my understanding according to the charter and when we discussed engagement with other standards organizations, my understanding was that this referred to liaising and coordination, rather than designing our work to accommodate their specific needs. For instance, we wouldn’t be expected to develop vocabulary that operates or must be implemented at the asset level, correct?

On Tue, Nov 4, 2025 at 1:40 PM Lila Bailey <[e-mail]<mailto:[e-mail]>> wrote:
Thank you both, Leonard and Sebastian.

This is very helpful context for those of us who joined the group more recently and didn't have the benefit of your briefing last year. I appreciate your patience and willingness to repeat yourselves.

All best,
Lila



On Mon, Nov 3, 2025 at 7:03 PM Sebastian Posth <[e-mail]<mailto:[e-mail]>> wrote:
Dear Lila,

Thank you for these important questions about other standards using the vocabulary.

TDM·AI as a Registry-Based Implementation

"You have mentioned other standards that plan to use this vocabulary, but it would be extremely helpful if you could tell us exactly what those standards are."

One example is TDM·AI (https://tdmai.org<https://tdmai.org/>) – a registry-based attachment mechanism that intends to use the IETF vocabulary. We presented this at the IAB Workshop on AI-CONTROL in Washington (September 2024) and have participated in the working group since then.

The approach is based on ISO 24138:2024 (ISCC)<https://www.iso.org/standard/77899.html> for content-derived identification and W3C Verifiable Credentials for cryptographic authentication of declarations.

TDM·AI Documentation

"It is impossible to understand the full impact of this vocabulary when the only context we have been looking at here in the IETF is robots.txt."

We provide full documentation at tdmai.org<https://tdmai.org/> that explains the registry-based approach in detail. TDMAI adopts the vocabulary's categories and definitions as a reference point to illustrate how domain-based attachment (as proposed by the IETF) can be translated into a registry-based system. TDM·AI demonstrates how preference declarations can be persistently and verifiably associated with individual digital assets via a registry, rather than relying solely on domain-based signaling like robots.txt.

Scope of Adoption

"I am also wondering if those other standards would be adopting the entire document, or just importing the categories and definitions?"

The Internet Draft defines "a vocabulary for expressing preferences regarding how digital assets are used by automated processing systems. This vocabulary allows for the declaration of restrictions or permissions for use of digital assets by such systems."

TDM·AI intends to adopt the entire vocabulary document – importing the categories and definitions and expressing them in digitally signed metadata declarations, which are externally bound to the ISCC fingerprint for anyone to resolve.

As I've mentioned on the list and in meetings, in my understanding this vocabulary is supposed to establish common ground for communicating AI preferences. Declaring parties and AI model or system providers need to understand each other – the vocabulary should provide that shared language.

Current Status

You will recognise, TDM·AI currently aligns with draft-ietf-aipref-vocab-02 (July 2025) but recognizes that both the IETF vocabulary and our implementation remain works in progress, subject to change! As the IETF draft evolves, we will update our alignment accordingly. Our documentation should be regarded as preliminary guidance showing one possible application of the vocabulary, not a finalized standard.

Hope that helps.

Best regards,
Sebastian

--
Sebastian Posth
Liccium.com


On Tue, 4 Nov 2025 at 01:55, Lila Bailey <[e-mail]<mailto:[e-mail]>> wrote:
Leonard-

You have mentioned other standards that plan to use this vocabulary, but it would be extremely helpful if you could tell us exactly what those standards are. It is impossible to understand the full impact of this vocabulary when the only context we have been looking at here in the IETF is robots.txt. I am also wondering if those other standards would be adopting the entire document, or just importing the categories and definitions? The fact that other standards are going to use this makes it all the more important that we understand the full scope of all types of uses and users that will be impacted, and ensure that these standards won't have negative/unintended consequences for under-resourced or vulnerable populations.

Thanks.

-Lila
--
ai-control mailing list -- [e-mail]<mailto:[e-mail]>
To unsubscribe send an email to [e-mail]<mailto:[e-mail]>
--
ai-control mailing list -- [e-mail]<mailto:[e-mail]>
To unsubscribe send an email to [e-mail]<mailto:[e-mail]>