Re: [dispatch] [Secdispatch] [art] Open Ethics Transparency Protocol

[Nikita Lukianets <[e-mail]>]

Hi everyone,

Related to the disclosure formats, we are adopting Model Cards and Data Sheets to a simple JSON file hierarchy which could be parsed and checked and extended using any programming language. 

The core idea behind is making machine-readable and human-readable labels interoperable and trusted thanks to a chained validation process. Simple visual example for it exists in creative commons generators though has no machine-readable component. 

I apologize for not responding to the comments. Some of you may know that I'm currently in Ukraine and things are, how could I say it... "complex". I don't have reliable internet connection these last days, especially when we have to stay at the bomb shelters. 

Indeed, Application layer seems most viable option here and I hope those interested could join the session in art-dispatch, while I'll try to make sure I address previously mentioned comments, especially those related to implementation and examples. 

Best, 

Nikita Lukianets 

On Mar 14, 2022 19:47, Kathleen Moriarty <[e-mail]> wrote:

Greetings!

I agree with the prior assessment that this is similar to the Software Bill of Materials work, at least in objectives. That is flexible in terms of what formats will be acceptable (3) to provide a manifest on software and many have selected SPDX, then it's digitally signed (code signing). There's another proposal for this upcoming meeting that will link these for use in supply chain assurance using Remote Attestation. It might be worth reviewing that proposal to look for similarities in addition to reviewing the document from the NTIA on the minimum set of elements for an SBOM.

There's a requirement in the US to any organization that sells to the federal government to provide an SBOM related to the US executive order from May 2021.  As such, there's traction on this approach already. With that said, EKR also pointed out the proposal in this thread is at the application layer. There may be a possibility for your work to use the standards being adopted that could accelerate your progress. The next part would be to determine if that work requires adoption and further standardization or if the existing formats suffice for what needs to be interoperable.

At this point, I don't see a time slot being useful, but please respond with your thoughts if I'm missing something. I hope this is helpful.

Thank you,

Kathleen

co-chair of SecDispatch

On Sun, Mar 13, 2022 at 9:39 PM Dale R. Worley <[e-mail]> wrote:

[e-mail] writes:
> Nikita Lukianets from the Open Ethics initiative here.
>
> I've been working on the mechanisms to enable transparency for data
> collection and data processing practices for autonomous systems and
> specifically, those powered by machine learning models. Since 2020 I
> have started to draft a guiding document to reflect ways disclosures
> could be submitted, verified, and exchanged. Eventually, I would like to
> see how this work could result in an open standard.

I have feelings which are similar to other respondents.  I phrase it
that the IETF is not the correct place for this work because the central
problem is at the application layer (and possibly above that, at the
political layer ... and certainly above that, at the cultural layer):
being transparent presupposes a suitable way to describe "data
collection and data processing practices" in "disclosures".

Once somebody defines a way to represent these disclosures as concrete
document objects, then there may be some protocol issues regarding how
to "submit, verify, and exchange" them.  But that still seems to be more
like the presentation layer, how you associate disclosure documents with
web sites etc. which they describe, which sounds like a W3C specialty.

Dale

_______________________________________________
Secdispatch mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/secdispatch" rel="nofollow">https://www.ietf.org/mailman/listinfo/secdispatch

--

Best regards,

Kathleen