"Authentic Web" Workshop

[dontcallmedom]

Originator Profile and C2PA (Coalition for Content Provenance and Authenticity) specifications provide distinct mechanisms to annotate Web content with provenance information, with a common goal of helping users assess the authenticity of digital materials they are finding. Both intends to be applicable to varying degrees to Web browsing.

Both were presented and discussed during TPAC 2024 breakouts:

• https://www.w3.org/2024/09/25-trust-origin-minutes.html
• https://www.w3.org/2024/09/25-authentic-web-minutes.html

From these discussions, it seems likely that a W3C Workshop on the topic might help identify common goals, realistic Web platform primitives and integration points that would help have a material impact on the challenges end-users are facing with regard to mis- and dis-information on the Web. The Technological Approaches to Improving Credibility Assessment on the Web report produced by the Credible Web Community Group likely provides a useful structure to have such a conversation.

With that said, it is also clear from the breakout discussions that more work is needed in focusing the topics on which W3C is the most likely to serve as an impactful convener.

[lrosenthol]

@torgo and I have been trying to advance the relationship between C2PA and the W3C, so anything that either/both of us can do to make this happen - and more important have it continue into actual reviews and technical work - let us know!

[dontcallmedom]

see also:

• a request for Mozilla's position on C2PA (empty at the moment)
• a draft proposal is exploring a possible browser integration of reputation management

[TzviyaSiegman]

See also https://misinfocon.com/about - articles and proposals for tooling to prevent spread of misinformation. Authors include current and former chairs of CredWeb Scott Yates and Connie Moon Sehat

[ylafon]

see also:

• a request for Mozilla's position on C2PA (empty at the moment)

• a request for WebKit's position on C2PA

[plehegar]

cc @chrisn

[wareid]

This discussion is needed, and I see the importance of focusing it a bit as we could potentially talk for ages about an infinite number of things.

This topic is very important to the digital publishing industry, and I think finding a solution or examining the use cases in that industry would be helpful to the workshop. Anti-counterfeiting of content for example.

[RByers]

I'd love to see the W3C tackle this space from a principles perspective. Perhaps something like 'what properties would trust systems need in order to align with the W3C Ethical Web Principles'? And perhaps "How should the W3C evaluate the benefits and harms caused by any specific trust framework"?

@jyasskin and I have been exploring some ideas here in the context of the Chromium project. I'd certainly be interested in input from the W3C community on what properties would be valuable to have in any such reputation system in Chromium. If there was a workshop on this somewhere, I would personally try to attend in-person.

This is one of the most important questions facing the future of the web, and in some way reflects the fundamental tension between openness and safety which the web has historically navigated so well.

[shigeya]

Hi. GitHub doesn't tell us a discussion of this kind is going on here. Some of my colleagues found this and informed me (as the Originator Profile's technological lead.)

Anyway, I hope to have a workshop soon, hopefully in springtime.

by the way,

With that said, it is also clear from the breakout discussions that more work is needed in focusing the topics on which W3C is the most likely to serve as an impactful convener.

Note that we are creating multiple specifications; some of them are for W3C, and the others are for IETF.

[csarven]

I'd also be interested in a workshop like this, as it aligns with the work we're doing on the dokieli project: https://dokie.li/

[dontcallmedom]

based on input received by email, @TzviyaSiegman and I will be trying to figure out how to articulate a possible series of Credible Web CG meetings possibly followed by a more in-depth synchronous workshop once these meetings have helped identify a proper structure to the discussions; we'll update this issue as make progress on that plan

[dontcallmedom]

@martinthomson 's thoughts on the topics (with C2PA as specific point of reference) https://lowentropy.net/posts/c2pa/

[chrisn]

@martinthomson 's thoughts on the topics (with C2PA as specific point of reference) https://lowentropy.net/posts/c2pa/

"A simple method would be to have content signed by a site that claims it. That immediately helps with the problem of people attempting to pass fake information as coming from a particular source."

(At the risk of turning this issue into a discussion, which should probably really happen in CredWebCG...) The above is really the problem that the BBC is interested in solving. It's not about claiming what's truthful or not, just being able to reliably establish where something came from (where traceability to BBC as publisher is far more important than traceability to a particular camera or editing software). Don't look at C2PA in isolation, need to also look at Creator Assertions Working Group and IPTC Origin Verified News Publisher Lists

[TzviyaSiegman]

From the scholarly publishing world https://zenodo.org/records/14107216 (research about the need to detect falsified images and/or certify that images are "authentic"