Re: "A Suggestion for Bob" (Moved to public list)

While I argue that an extended variant of W3C Web Annotations is the
appropriate way to support the publication and discovery of 2nd and 3rd
party credibility signals (i.e. not 1st party signals embedded in the
content itself), I recognize that just about any mechanism for stating a
relationship between a web resource and some other retrievable resource
containing a credibility signal would satisfy the technical requirement.
This means, of course, that anything that supports RDF or an RDF-like
mechanism would do the job. In fact, it is useful to observe that the
ability to assess credibility would be enhanced if there is access not only
to explicit credibility signals but also to the full range of data and
relationships that are recorded in the Semantic Web. But, while the claim
"All you need is RDF!" has a great deal of truth to it, I feel that claim
is insufficient in that it addresses only one aspect of the problem set;
data storage conventions. "Just use RDF," tells us something about how to
store the relationship data, but it says nothing about the expected use
cases, interfaces, dialog patterns, etc. whose understanding is essential
to the development of a useful communication paradigm. (Credibility signals
are communications...) "Just use RDF," does little to allow us to build up
a common, shared understanding of how these systems might "look." On the
other hand, speaking about W3C Web Annotations invokes a particular set of
assumptions, understandings, etc. of one very useful means of establishing
relationships between resources.

When I pull up a web page in my browser (say: "
https://example.com/post1.html") I would like my browser to present to me,
in parallel with the page, a filtered/sorted list of annotations relevant
to that page. Determinations of relevance should be based on the current
state of my personal "Credibility Web" (see Sandro's earlier posts re
Credibility Web). If there are no annotations in the curated set, or if I
am curious to see what is not in the curated list, I should also be able to
request either a full set of annotations, or a set filtered and sorted
using something other than my default Credibility Web. Ideally, the browser
would also be able to pull in and display, in some useful format, data
concerning the page, its hosting site, author, etc. that is retrieved from
the Semantic Web. Clearly, it would be useful if processes in my browser,
or that work in cooperation with my browser, could evalature the same set
of annotations or Semantic Web data and generate various summaries,
"scores," etc. that could be presented to me in order to enrich my ability
to assess the credibility or truth of claims made on the current page.

Given that W3C Web Annotations are "just RDF or RDF-like stuff," I could
say "Just use RDF," but I think it is more useful to view this from the
users point of view, to see it as an application of the annotation use
case, and leave the use of, or necessity for, RDF as a technical
implementation detail.

bob wyman


On Sun, Sep 5, 2021 at 1:15 PM Bob Wyman <bob@wyman.us> wrote:

> The message exchange reconstructed below occurred on the internal list,
> but it has been observed that it belonged on the public list instead. I
> hope the other contributors to this thread do not object to my repackaging
> and reposting it on the public list.
>
> ----- Forwarded Messages follow -----
>
> From: Owen Ambur <Owen.Ambur@verizon.net>
> Date: Sun, Sep 5, 2021 at 12:14 AM
> Subject: A Suggestion for Bob
> To: <internal-credibility@w3.org>
>
> Sandro, thank for sharing these audio/video recordings.
>
> Since Bob Wyman is troubled by the lack of capability to express to the
> world that he is not a communist, the purpose of this message is to offer
> him a potential means of relief:
>
> Publish his plan(s) on the Web in StratML Part 1, Strategic Plan (ISO
> 17469-1) format.
>
> If he were to do so, it would be fairly easy for myriad AI-enabled
> intermediary services to pretty well establish whether he is a communist or
> not, assuming of course that he honestly documents his vision, mission,
> values, goals, objectives, and stakeholders -- as best he understands them.
>
> With reference to Leonard Rosenthol's comment about verification, that
> depends upon whether performance indicators of actual results are reliably
> documented and shared -- preferably in records published in an open,
> standard, machine-readable format like StratML Part 2, Performance Plans &
> Reports (formerly ANSI/AIIM 22:2017).
>
> Lacking such indicators anyone's guesses and "assertions" are as good as
> anyone else's.  Even if Bob does openly represent himself as a
> card-carrying communist, that alone doesn't truly make him one  For
> example, he could be working as an undercover agent for the FBI ... or a
> spy for fascists ... or a college professor with tenure in an ivory tower
> insulated from reality.  Or perhaps he may *mistakenly* believe that he
> is a communist.
>
> What matters is the degree to which he actually participates in
> commandeering other people's property, discouraging personal initiative,
> denying personal responsibility, destroying economic incentives, and making
> everyone poorer, i.e., what real communists do.
>
> With respect to your concluding comment, I look forward to learning
> whether we can do more together than merely "attend a few meetings and
> exchange E-mail messages".  Perhaps I'll have a better sense of that when
> I've had a chance to listen to recording of the meet-the-candidates
> meeting.  If so, I'll "scribe" my understanding the the group's plan in
> StratML format.
>
> Owen
> ---------------------------------
> From: Bob Wyman <bob@wyman.us>
> Date: Sun, Sep 5, 2021 at 1:33 AM
> Subject: Re: A Suggestion for Bob
> To: Owen Ambur <Owen.Ambur@verizon.net>
> Cc: <internal-credibility@w3.org>
>
>
> Owen,
> It would be nice if your "myriad AI-enabled intermediaries" could support
> Bob's denial of Alice's claim. However, one must wonder if the output of
> those intermediaries would actually be useful. The problem is that unless
> Alice is willing to allow claims of the intermediaries to be posted
> alongside her own claim, very few, if any, of the people who read Alice's
> claim will see any claims, rebuttals, or proofs that challenge Alice's
> credibility. I think this is a major problem with what appears to be the
> current thinking about these credibility signals. For credibility signals
> to be more than just a means for authors and their hosts to publish
> self-serving claims that support, but do not challenge, their own
> credibility, we need to describe how credibility signals, both positive and
> negative, can be associated with authors, statements, etc. without the
> consent of their subjects and potentially even without their knowledge
> (although such signals need not be hidden). This is why I've pointed to
> annotation protocols on several occasions. Credibility signals, published
> as Web Annotations, would have the needed characteristics.
>
> Of course, the existing W3C Web Annotation protocol is incomplete, for the
> desired purpose, since it doesn't define the "search" function needed to
> discover annotations that are related to a specific URL. Personally, I
> think it would make sense for this group to define Credibility Signals as
> annotations and to define the additional protocol needed to allow discovery
> of such annotations. If we were to do this, then your "myriad AI-enabled
> intermediaries" would be able to publish their assessment of Alice's claim
> in a manner that is more likely to be discovered by readers of Alice's
> claim -- if only people were to use annotation clients... Your AI-bots
> might even be enhanced to search out other posts with similar claims and
> annotate them as well.
>
> bob wyman
>
> ---------------------------------
> From: Georg Rehm <georg.rehm@dfki.de>
> Date: Sun, Sep 5, 2021 at 2:07 AM
> Subject: Re: A Suggestion for Bob
> To: Bob Wyman <bob@wyman.us>
> Cc: Owen Ambur <Owen.Ambur@verizon.net>, <internal-credibility@w3.org>
>
> Dear Bob,
>
> On 5. Sep 2021, at 07:33, Bob Wyman <bob@wyman.us> wrote:
>
> Owen,
> […]
>
> Of course, the existing W3C Web Annotation protocol is incomplete, for the
> desired purpose, since it doesn't define the "search" function needed to
> discover annotations that are related to a specific URL. Personally, I
> think it would make sense for this group to define Credibility Signals as
> annotations and to define the additional protocol needed to allow discovery
> of such annotations. If we were to do this, then your "myriad AI-enabled
> intermediaries" would be able to publish their assessment of Alice's claim
> in a manner that is more likely to be discovered by readers of Alice's
> claim -- if only people were to use annotation clients... Your AI-bots
> might even be enhanced to search out other posts with similar claims and
> annotate them as well.
>
>
> just as a follow-up to what you sketched in your email, this is an idea
> that I described in the following paper a number of years ago:
>
> Georg Rehm. “An Infrastructure for Empowering Internet Users to handle
> Fake News and other Online Media Phenomena”. In Georg Rehm and Thierry
> Declerck, editors, *Language Technologies for the Challenges of the
> Digital Age: 27th International Conference, GSCL 2017, Berlin, Germany,
> September 13-14, 2017*, Proceedings, number 10713 in Lecture Notes in
> Artificial Intelligence (LNAI), pages 216-231, Cham, Switzerland, 2018.
> Gesellschaft für Sprachtechnologie und Computerlinguistik e.V., Springer.
> 13/14 September 2017.
>
> https://link.springer.com/content/pdf/10.1007%2F978-3-319-73706-5_19.pdf
> <https://link.springer.com/content/pdf/10.1007/978-3-319-73706-5_19.pdf>
>
> The article proposes an infrastructure to address phenomena of modern
> online media production, circulation and manipula- tion by establishing a
> distributed architecture for automatic processing and human feedback. The
> main building blocks of the proposed infrastructure are: natively embedded
> into the World Wide Web, Web Annotations, Metadata Standards, Tools and
> Services, Decentralised Repositories and Tools, Aggregation of Annotations.
>
> Best regards,
> Georg
>
> ---------------------------------
> From: Bob Wyman <bob@wyman.us>
> Date: Sun, Sep 5, 2021 at 2:55 AM
> Subject: Re: A Suggestion for Bob
> To: Georg Rehm <georg.rehm@dfki.de>
> Cc: Owen Ambur <Owen.Ambur@verizon.net>, <internal-credibility@w3.org>
>
> Georg,
> Nice paper. I think you and I have a common vision of the solution space
> that needs to be explored. The question, of course, is whether or not the
> Credibility group is the correct context for these discussions. I am hoping
> that it is as I am unaware of other good candidates.
>
> Given the importance of having annotations supported in the browser, we
> might be encouraged by the fact that there is an existing effort to get
> Chrome and Chromium to provide support for annotations. (See:
> https://github.com/bokand/web-annotations ) However, I am concerned that
> if that effort moves ahead without substantive input on what must be done
> to ensure that annotations are useful for credibility signalling, it is
> entirely possible that we may find broad deployment of an annotation
> interface that makes it difficult to add in the later support at a later
> time. Also, there is some risk that a browser built-in annotation system
> might make assumptions about the storage of annotations that make it
> difficult to provide them via a federated system. We may get locked into
> monolithic annotation systems which would not address the needs we both
> appear to recognize.
>
> bob wyman
> (Resident of Zehlendorf and Dahlem during the 1960's... Berlin bleibt doch
> frei!)
>
> ---------------------------------
> From: Georg Rehm <georg.rehm@dfki.de>
> Date: Sun, Sep 5, 2021 at 4:36 AM
> Subject: Re: A Suggestion for Bob
> To: Bob Wyman <bob@wyman.us>
> Cc: Owen Ambur <Owen.Ambur@verizon.net>, <internal-credibility@w3.org>
>
> Bob,
>
> yes, the idea and vision I described in the paper relies on the massive
> adoption of web annotations (and their full implementation and availability
> in all browsers), on decentralised annotation repositories, on a
> decentralised set and toolbox of text mining and natural language
> processing tools that operate on the decentralised annotations and on
> shared vocabularies, i.e., annotation formats.
>
> Georg
>
> P.S. I live in Prenzlauer Berg, close to Mitte. Not immediately accessible
> to residents of Zehlendorf or Dahlem in the 1960s.
>
>
>
>
>