[MINUTES] VCs for Education 2025-09-15

kte-hamg-bpj Meeting Summary (2025-09-15)

*Topics Covered:*

   - *Challenges with Automated Meeting Invitations:* Ildiko discussed
   difficulties restarting automated meeting invitations after a summer break.
   - *Impact of Political Climate on University Interactions:* Phillip
   shared an anecdote about restrictions on non-faculty members interacting
   with students at UVA, highlighting concerns about access and collaboration.
   - *ISO Mobile Driving License Standards and Verifiable Credentials:* A
   significant portion of the meeting focused on the ISO mobile driver's
   license (MDL) standard, its verification methods (phone home vs. device
   verification), and its suitability for education use cases. Concerns were
   raised about the "phone home" verification method and its privacy
   implications. The group discussed the lack of a clear, universally accepted
   standard and the need for a definitive resource to point people to (W3C
   verifiable credentials were suggested).
   - *Suitability of Mobile Driver's Licenses for Education:* Simone
   highlighted the limitations of MDLs for representing the granularity and
   expressiveness needed for educational credentials, advocating for
   verifiable credentials instead. A vendor's unexpected switch from
   verifiable credentials to MDLs without customer awareness was cited as an
   example.
   - *Legal Landscape and Educational Technology:* Jeff noted the positive
   impact of recent changes in educational technology law. The group suggested
   inviting speakers to provide overviews of legal landscapes in different
   jurisdictions.
   - *Upcoming Conferences:* The Internet Identity Workshop (IAW) and EPIC
   (conference on open recognition) were mentioned.
   - *Data Models for Verifiable Credentials in Education:* Discussion
   included individual case data models, digital diplomas, course credentials,
   student IDs, and open badges.
   - *AI and LLM Applications in Credential Creation:* The use of LLMs and
   AI to assist in creating verifiable credentials was discussed, focusing on
   challenges and opportunities in automating credential description and
   criteria writing. The need for large, high-quality datasets for training
   these models was highlighted.
   - *Interoperability and APIs:* The group discussed interoperability
   initiatives, data portability between wallets, and the need for APIs to
   facilitate credential consumption by third parties. Updates on operating
   system vendor efforts in integrating verifiable credential APIs were also
   discussed.
   - *Skill Inference from AI:* The meeting discussed challenges and
   opportunities related to inferring skills from AI, including data sources
   (course descriptions vs. LMS data), and the accuracy and reliability of
   inferred skills.
   - *Proficiency Levels and Skill Standards:* The HR Open standard for
   interchanging proficiency levels by skill was mentioned.
   - *Multi-skill Credentials and Data Transmission to States:* The meeting
   included discussions on creating multi-skill credentials using OBV3 and on
   projects related to transmitting structured data to states for employment
   recording purposes.
   - *Terms of Use for Credentials:* The group discussed the need for terms
   of use for credentials, allowing issuers and holders to specify data usage
   restrictions.
   - *Learning and Work Encyclopedia Resources:* The group briefly
   discussed a resource compilation by the Learning and Work Encyclopedia.
   - *Open Badge Extensions:* An upcoming session at the ATEMPPU conference
   focused on developing open badge extensions aligned with favorable
   properties was mentioned.

*Key Points:*

   - There's a significant need for a clear, widely accepted standard for
   verifiable credentials in education, with W3C standards being a potential
   solution.
   - The privacy implications of various verification methods for
   credentials, particularly the "phone home" approach, are a major concern.
   - The limitations of existing standards, like the ISO MDL standard, for
   the nuanced needs of educational credentials were highlighted.
   - The potential of AI and LLMs to streamline credential creation is
   promising, but high-quality training data is crucial.
   - Interoperability and API development are essential for wider adoption
   of verifiable credentials.
   - The accurate inference of skills from data sources remains a
   challenge, requiring careful consideration of data sources and validation
   methods.
   - Terms of use are necessary to clarify data usage permissions for
   credentials.

Text:
https://meet.w3c-ccg.org/archives/w3c-ccg-vcs-for-education-2025-09-15.md

Video:
https://meet.w3c-ccg.org/archives/w3c-ccg-vcs-for-education-2025-09-15.mp4
*kte-hamg-bpj (2025-09-15 11:00 GMT-4) - Transcript* *Attendees*

Alex Higuera, David Ward, Dmitri Zagidulin, Eric Shepherd, Hiroyuki Sano,
Ildiko Mazar, JeffO - HumanOS, Kayode Ezike, Naomi Szekeres, Phil Barker,
Phillip Long, Simone Ravaioli, Ted Thibodeau Jr
*Transcript*

Ildiko Mazar: Hello. Sometimes where are you?

Simone Ravaioli: Italy finance.

Ildiko Mazar: What? Yeah.

Simone Ravaioli: Back to school today in Italy.

Ildiko Mazar: For how long? A week and then you're off again.

Simone Ravaioli: I might see you at the one tech EU thing. So that's a
week, right? There's

Ildiko Mazar: That's what I thought.

Ildiko Mazar: Okay, And Deitri, Welcome back.

Dmitri Zagidulin: Hi, Elico. Hi, Simony. Hi, Ted. Yeah, guys, it's good to
be back. It's exciting.

Ildiko Mazar: It's been a long summer. Very hot here.

Dmitri Zagidulin: Yeah.

Ildiko Mazar: Yeah. Catalonia is notorious for very hot summers and it's
only getting worse more time.

Dmitri Zagidulin: I was going to say, isn't it drought a bit?

Ildiko Mazar: Mhm. Yeah.

Ildiko Mazar: Although we had some good rains this year, so at least we
don't have the restrictions. There was a time when we were restricted to I
don't know many liters per household.

Dmitri Zagidulin: Okay. …

Ildiko Mazar: They didn't even consider the number of people in the
household.

Dmitri Zagidulin: Wow.

Ildiko Mazar: He just said, I don't know, 42 liters. That's it.

Dmitri Zagidulin: That's intense.

Ildiko Mazar: So maybe we should reconsider.

Dmitri Zagidulin: How did they just out of sheer curiosity, how did they
enforce that? Was it fines afterwards or did they just not serve the water?

Ildiko Mazar: We were good. So we don't know how others were find or
otherwise, but they do have records. So they sent letters I know that much
but whether there were any fines or punishments of any sort don't know Nomi
hello okay people are coming good I wasn't entirely sure…

Dmitri Zagidulin: Absolutely. Yeah.

Ildiko Mazar: if I managed to restart the automatic notifications because I
have a weird

Ildiko Mazar: as to the calls. I can change some things and not other
things. And it all started happening with a predated invitation for an
entire series. So, it's good to see that people are coming.

Naomi Szekeres: Hi. Yes.

Dmitri Zagidulin: Yeah. No worries. I'm certain we'll sort it out.

Naomi Szekeres: The send out of can you hear me?

Dmitri Zagidulin: Go ahead, Naomi. Yes.

Ildiko Mazar: And you received it.

Naomi Szekeres: I was just going to say the sendout of the invite that we
could post on our calendars was very helpful.

Dmitri Zagidulin: Go ahead, Naomi.

Naomi Szekeres: Yes, I did and I was able to download it without a problem
and…

Ildiko Mazar: Okay. …

Naomi Szekeres: it posted to Not a problem. Thank you.

Ildiko Mazar: it's I don't know…

Simone Ravaioli: This hurts.

Ildiko Mazar: who set up the original one. I just stopped it for the summer
and I was crossing fingers that I could restart it last week because it's
something that you can only do after the last cancelled date and then it's
automatic.

Naomi Szekeres: Got it.

Ildiko Mazar: Tell hello.

Phillip Long: Goodbye.

Dmitri Zagidulin: And I'm sure Hi, hi, Phil. I'm sure folks will need a
couple weeks to get used to get this in the back of their minds and so on.

Ildiko Mazar: Yeah. I don't know…

Dmitri Zagidulin: We'll sort it out.

Ildiko Mazar: how the other groups are doing it. It's been a very busy
summer for the rest of

Dmitri Zagidulin: Yeah, I think other groups don't really take breaks like
that. Although I suppose no,…

Dmitri Zagidulin: I suppose there is precedent. so the answer is with
difficulty with exactly the same as we're doing right now. nice.

Simone Ravaioli: I'd love to break records…

Simone Ravaioli: if it is about unplugging. Curve

Ildiko Mazar: Yes.

Dmitri Zagidulin: Yeah. Yeah. Yeah. Yeah. we're spoiled in education.
because it's sort of traditional, right?

Eric Shepherd: Hello. Sorry.

Phillip Long: It's funny. I just got back from a meeting on the UVA grounds
talking with a guy who runs the called project. It's a synthetic biology
team project

Ildiko Mazar: what?

Phillip Long: which I helped start at MIT back in 2003 and going worldwide
now. But I learned that in the current political climate people formerly
members of the faculty at UVA are not permitted to be in a room with
students.
00:05:00

Simone Ravaioli: Okay.

Dmitri Zagidulin: Whoa.

Phillip Long: If you're not a vetted member of the faculty staff at the
institution then you cannot sit in a seminar with students or an
extracurricular project or anything like that. It's prohibited or the
faculty adviser will lose their position.

Eric Shepherd: Is that the same for virtual events as well as physical
events? What's

Dmitri Zagidulin: That's intense.

Phillip Long: It's not the same for virtual events if it is advertised to
the faculty but because at many institutions and UVA is among them there is
a login that vets you before you get into the event…

Ildiko Mazar: No heat.

Phillip Long: if you're not a member of the organization. So if they make
it intentionally public then you can do it. they just will not let you on
grounds and sit in a seminar class to have a discussion with them or…

Simone Ravaioli: Oops.

Phillip Long: in this case it's an extracurricular team that meets on its
own to decide how to genetically engineer a new protein other biologically
active entity to try to solve a problem. and…

Eric Shepherd: You've only just noticed, right?

Phillip Long: it was very disappointing and kind of made me worry…

Ildiko Mazar: Boy, you're

Phillip Long: where our institutions are going if the city on the hill now
has a moat around it. it's tough for interactions with others that are not
part of that select group.

Simone Ravaioli: Stop it.

Phillip Long: I guess it hadn't been so much in my face until that
particular conversation because that's not the way it started. you'll see…

Eric Shepherd: I have a question for the group and…

Phillip Long: if you go to igim.org you'll see there are 400 teams from
around the world that participate in it. so all countries everywhere

Eric Shepherd: kind of someone could point me in the right direction
because I see that the ISO published the mobile driving license and…

Phillip Long: Yep. You want to start?

Eric Shepherd: also there's another on taking mobile driving license IDs
online and I'm turning up to lots of discussions around verifiable
credentials…

Eric Shepherd: but is there going to be an ISO standard or is there going
to be a standard based on those ISO standards or are we all just kind of
enjoying conversations but not end up with a definitive standard that this
is

Dmitri Zagidulin: Great question.

Dmitri Zagidulin: So, yeah, let me start and then everybody feel free to
join in. So, there's a couple things going on here. So on the W3C side
there is verifiable credentials 2.0 model. So that's a global standard. we
are just now rechartering the next iteration of the WC verifiable
credential working group. So BCA model 2.1 but you asked about ISO
specifically and mobile driver's license.

Dmitri Zagidulin: So mobile drug go ahead.

Eric Shepherd: And…

Phillip Long: Go ahead.

Eric Shepherd: there. Sorry,…

Phillip Long: You froze for a minute.

Eric Shepherd: You froze. I didn't know where you were thinking or…

Dmitri Zagidulin: No, no.

Eric Shepherd: when you froze. Mhm.

Dmitri Zagidulin: I froze mentally, not technologically. I'm trying to
think of how to phrase it. So, a couple things are going on, So we have so
ISO put out the mobile driver's license standard and the MDOC standard
which is kind of the parent class of so mobile driver license is an example
of a mobile document and for better or for worse it's an ISIL standard but
how do we put it there's a lot of industry wrangling there's a lot of
monopolist lobbying

Dmitri Zagidulin: of various government and educational structures. there's
tension between what's in the core spec versus the proprietary components
that the implementers bring to it. So the whole mobile driver's license is
complicated topic. In addition will ISO handle DIDs and verifiable
credentials? I hear rumors something to that effect,…

Dmitri Zagidulin: but I don't know much detail. I don't know if Simony or
Ilic or anyone else has more insight.

Phillip Long: So, I'll chime in and…

Phillip Long: say that at the last IIW there was major discussion about
this with some degree of surprise because people have apparently hadn't
been paying attention to the details of it.
00:10:00

Dmitri Zagidulin: yeah.

Phillip Long: And the surprise is that one of the verification methods that
is recommended is that the credential is verified by quote unquote phoning
home to the issuer. and that is effectively a surveillance method because
you can include other kinds of information in the credential itself.

Simone Ravaioli: Oops.

Phillip Long: If the query is so formed when that is verification is done
which if you go to no phonehome.com you'll see a statement about it and
organizations that have expressed phonehome.com their concern for that
approach partly because you can there the other method for verification is
called device verification as opposed to server verif ification, which is
the phone home side. And device verification relies on the person who is
doing the verification making the request to an independent verifier or
issuer service and not necessarily anything to do with the issuer
themselves. and unfortunately that's not the recommended approach.

Phillip Long: It's also not the most efficient approach in the way it's
been implemented and so people are attempting to implement that one. But in
California for example, they have implemented the MDL using OIDC and
verifiable credential components to do so. you can implement MDLs in the
sense of a mobile driver's license other than using the ISO standard that
came out of this particular 13 whatever it is version of it 13. Yeah.

Simone Ravaioli: Awesome. Come on.

Eric Shepherd: 18. Yeah.

Phillip Long: And in fact to speak to Dimmitri's comment because of the
push back that's emerged we're told that they have pulled that
recommendation of a serverbased verification out of the core standard but
all they have done is move it to a separate document so that it could be
implemented with their guidance and…

Simone Ravaioli: It's a

Phillip Long:

Phillip Long: and that's because This is my view and I think it's a view of
others. ISO is an organization that is the voting membership is the
countries that are members and there is a country representative and has
the vote and there are more autocratic and other centralized economies that
are very much interested in having the government be the one that looks at
the required verification as opposed to anything else. And so that's the
way it's been implemented.

Eric Shepherd: I participate in the learning and tech education technology
the SC36 group, and that's not what I find in that group. I mean, they are
all genuinely interested in doing standards that will be internationally
accepted. The advantage of ISO is that governments find them easy to adopt.
and…

Phillip Long: Absolutely. That's right.

Eric Shepherd: they might be slow, but they've got two standards out there
that feel like they're very appropriate and they'll probably be adopted by
governments because it's a safe place to go. Feels like when people ask me,…

Simone Ravaioli: let's

Eric Shepherd: which one should we use? we're not talking about
international standards. We don't have anything. But maybe I'm missing it.
That's why I asked the question, it's like which standard should I say?
This is the standard. This is what it's all going to be based on. And as
opposed to I'm not qualified or competent or intelligent enough to have a
debate about the options. It's like, tell me the URL, that's the one I'll
point people to.

Phillip Long: Right. …

Eric Shepherd: And it feels like we should be at that stage,…

Phillip Long: ISO does have international in its name.

Eric Shepherd: Yeah. So I should

Phillip Long: I would argue that W3C is an international standards
organization as well.

Simone Ravaioli: Awesome.

Phillip Long: As there are others. and so you have I guess the joke in the
standards world is the wonder thing about standards that's so wonderful is
that there are so many of them.

Eric Shepherd: Mhm.

Phillip Long: But in this particular case, the fact that the requesttor can
ad hoc choose to do the verification based on a perverification event to do
it through the phone home method or do it through a device method without
telling the individual that this is going on and without giving the
individual the option to say I would prefer not to have it done that way.
that's the concern that some of us have about this approach. and there's
nothing on your screen other than that says you're getting your credential
for your driver's license is being sent to the verifier But at any It could
be done in one way or another and you would have no way of knowing.
00:15:00

Phillip Long: And you also have no way of knowing if they've added
additional queries to the verification process which might include things
like your geol location at the time the query is made or the verification
is made so that they can know where you are. In the US that's particularly
an issue because we use our driver's license for more than driving. we use
it when we get prescriptions. We use it when we enter restricted places for
alcohol bars and such. We use it when we're buying certain types of
products that have an age verification associated with them. All of which
if it is done by verification with your driver's license is conveying to
people what you are acquiring and when.

Phillip Long: which may not be the most useful thing to have in
environments…

Simone Ravaioli: Okay.

Phillip Long: where you might be getting a tele via a prescription provider
in Massachusetts for the termination of an abortion and…

Phillip Long: you happen to be living in Texas.

Eric Shepherd: I mean,…

Eric Shepherd: there are already wallets out there with electronic driving
licenses.

Eric Shepherd: So, Colorado does it. I don't know which standards they've
based their stuff on.

Phillip Long: Yeah, there's about 15 at least at the moment,…

Phillip Long: maybe more. And six or seven of them use verifiable
credentials and the others use the ISO. There's a meeting happening in Utah
on the 15th and 16th I believe it is of October that the governor of Utah
has called to talk about the privacy concerns of the different methods for
doing driver's license because the state of Utah implemented the ISO
standard with recommend the recommended verification

Phillip Long: being the server verification. And when it was pointed out by
their chief privacy officer for the state that this was going on, the
governor requested and the parties got together and drafted a bill that
said they will not do server verification in Utah for their driver's
licenses. and they passed it.

Eric Shepherd: Yeah, I didn't Yeah,…

Eric Shepherd: I didn't want to hijack the meeting, but I think it's

Eric Shepherd: very illustrative of the confusion. If we're not careful,
we're going to be debating this for 10 years and some other, standard that
we disagree with is going to become the standard. So, it'd be nice to
actually get something to refer people to. And it sounds like W3C is the
one I should refer people to.

Dmitri Zagidulin: Yes. Yes.

Eric Shepherd: Yeah. Yeah.

Dmitri Zagidulin: That's Yeah,…

Phillip Long: There is a resources page on nofphone.

Dmitri Zagidulin: that's basically the case.

Eric Shepherd: Thank you.

Simone Ravaioli: Thank you.

Phillip Long: That also gives you more information.

Eric Shepherd:

Eric Shepherd: Thank you. How loud?

Dmitri Zagidulin: Simony, you have your hand up.

Simone Ravaioli: This was exactly one of the topic that we wanted to bring
to this community.

Simone Ravaioli: There's a lot going on and some of it is down the rabbit
hole.

Simone Ravaioli: But if I try to make this issue understandable to an
audience of people that are in education mostly what I have been saying is
that I mean apparently from the way the standard is designed in education
we need a level of expressiveness and…

Eric Shepherd: It is.

Simone Ravaioli: granularity when we talk about the payloads and the way we
define skills and so on that is not well supported by a mobile driver
license which is more flat if you will. So an education use case is not
best served by a mobile driver license.

Simone Ravaioli: that could be enough for some stakeholders to decide which
way they're going to go or understand what's happening because I have been
personally living through some real cases this summer…

Eric Shepherd: All right.

Simone Ravaioli: where a vendor decided to switch from verifiable
credentials to MDL overnight and the customer was completely unaware of
what that meant but they were not really irmed by the vendor that they were
switching to mobile driver license where all they knew was this stuff goes
in a digital wallet.
00:20:00

Simone Ravaioli: So why some of the technical details are not as meaningful
for education customers when you start to talk about what do I want you
this credentials to represent deep skills data connections to jobs and the
stuff that we mostly talk about you I guess what we're saying is that the
work that we've done in the education space around digital credential
standards is best served by verifiable credentials or you call it OB3 and
so on, you lose some of that expressiveness if you look at mobile driver
license.

Simone Ravaioli: So I mean that's what I have understood by doing a little
bit of research but without really going into too much technical details.

Dmitri Zagidulin: Makes sense.

Dmitri Zagidulin: Makes sense. So, this is a good segue into our call topic
in general which is I suppose we should probably go through the boiler
plate. but everybody on here I think all familiar faces Anyone can join
these meetings. You don't have to be a member of the W3C or the community
group in order to participate. Though you do need to be a member if you
want to make substantive contributions to any of the work items that we're
working on. the calls are recorded. We are using I believe Gemini for
notetaking or something.

Dmitri Zagidulin: we're using some sort of AI transcription service. we
usually do introductions and reintroductions which I suppose we should give
an opportunity to anyone who's either new to the call or who hasn't had a
chance to introduce themselves in a while to just raise your hand and
introduce yourself to the group. we'll do any quick announcements and
reminders and then we'll launch into our main discussion which we've
already been doing right which is items of interest for this basically
topics for upcoming speed running through it. anybody wants to reintroduce
themselves feel free to raise your hand here on Google Meet.

Dmitri Zagidulin: I know the icon is in different places for everyone
versus mobile or desktop, etc. And they keep moving it, but assuming you
find it.

Dmitri Zagidulin: Aha, Jeff, go ahead.

JeffO - HumanOS: Hey there. I just want to say it's wonderful to see you
all back in your group here.

JeffO - HumanOS: Very very grateful for that. Also just wanted to observe
that I'm grateful for the change in school technologies law that has been
percolating over the last few weeks.

Simone Ravaioli: Yeah. responsible.

JeffO - HumanOS: It's been very very interesting and really spritly
apparently to the upside almost across the board. So good to see you all
back. A lot has happened and I'm grateful that you guys are here with us

Dmitri Zagidulin: Thank you,

Dmitri Zagidulin: And you do bring up a reminder that I wonder if it would
be possible to invite speakers to give us an overview of what's happening
in the legal landscape in the EU,…

Simone Ravaioli: Thank you.

Simone Ravaioli: Peace.

Dmitri Zagidulin: in the other jurisdictions, that potentially influence
what we're doing. Anyone else? announcements and reminders. classic
announcement that Internet identity workshop or IAW is coming up in the
fall. It's the week before Halloween in the US in Mountain View,
California. A lot of identity and verifiable credentials work is done there.

Dmitri Zagidulin: Of course, it's not a specifically education focused u
conference …

Dmitri Zagidulin: but it is of interest for those of you who go and others.
Yeah, go ahead.

Phillip Long: Did just you said a week before it is a two weeks before this
year…

Phillip Long: because of a scheduling snafu at the computer science museum.
So instead of the week leading into the Halloween weekend,…

Phillip Long: it is the week prior.

Dmitri Zagidulin: And that right there is…

Dmitri Zagidulin: why we need working groups because pinning down exact
definitions is hard. that is…

Dmitri Zagidulin: what I mean by the week before. but you have a very valid
and probably more correct interpretation of what the term the week before
means.

Dmitri Zagidulin: So yeah, basically it's the week of the 20th and the 24th
is IW.
00:25:00

Phillip Long: just being here.

Dmitri Zagidulin: Yeah. No, I appreciate that. and preciseness and
terminology is what we're all about here. So, I appreciate you mentioning
it. if anybody else knows of other conferences and…

Phillip Long: Yes. …

Dmitri Zagidulin: community announcements, please post them into chat.
We'll add them to the notes.

Phillip Long: Epic is this December, I believe, in Paris, the conference on
open recognition.

Ildiko Mazar: It's October,…

Phillip Long: And I'll look for the dates here in a second.

Ildiko Mazar: I believe. U but I was also searching for too many links at
the same

Phillip Long: Oober. Okay. Sorry.

Dmitri Zagidulin: Sounds good. so one of the things that I wanted to put
together is just a document we can all sort of edit we're going to pull the
group and say hey add your topics of interest things you would like to see
us present on raise your hand and we'll capture them here in this document.
so I'll be editing it, but please everybody else feel free to edit it as
well. the link to it is in a Google Meet chat. You don't need to have an
account in order to edit it. this is an initial brainstorm of some of the
topics that we could discuss, but I'd love this to be a collaborative
discussion.

Dmitri Zagidulin:

Simone Ravaioli: I need It's not

Dmitri Zagidulin: And thank you for adding the attendees. I'm assuming
that's Ted, thank you. all right.

Dmitri Zagidulin: So, Phil added a link to the epic in Paris.

Phillip Long: I'll put it in.

Phillip Long: I just put the text in. Hold on.

Dmitri Zagidulin: Brilliant. All let me sort of go over and introduce some
of these topics that we'll be talking about in the coming weeks. So as
always we want show and tell from our members. Everybody a lot of you are
working on cool stuff standards open source software products we want to
hear from you.

Simone Ravaioli: Yes.

Dmitri Zagidulin: So reach out to the chairs to Ilico Simony or…

Dmitri Zagidulin: myself and say hey I would love to present on X and we'll
fill you in. I think given that we're verifiable credentials in education,
I think there's still a lot of interesting discussion to be had on
individual case data models. a call in diplomas to see how the current
digital diplomas are being deployed, which schemas and…

Simone Ravaioli: Yes. Yes. That's

Dmitri Zagidulin: data models people are using, course credentials, student
ids, open badges, version three and so on as always displaying the
credentials, especially displaying in a more or less interoperable way.
wallets, ver verifiers, any sort of consuming software. very useful. Also,
do we have an echo? if anyone's echoing, we want to mute. I think Simony,
that might be There's a slight echo coming in from you. all right. What
else? so there's been a lot of interesting projects that are using LLMs and
AI in general to assist in the creation of verifiable credentials.

Dmitri Zagidulin: there is the eternal topic of hey cred credentials
especially achievements and course completion are always better with the
addition of when it's coded against scaletologies achievement anthologies,
right? we want to be able to have these credentials be machine actionable.
I like that term. and doing so is hard and process intensive.

Dmitri Zagidulin: And so a number of projects have been using machine
learning to be able to say here's the free form text description of this
badge and here are the skills as coded by I don't know the US Department of
Labor or here are these skills as coded by the anthology used by European
Union learning community etc. Right? So that there's also been a couple of
assisted basically AI help for issuers, right?
00:30:00

Dmitri Zagidulin: So, one of the minor but interesting insights for me over
the summer from the Badge Summit in Colorado I've met multiple people whose
full-time job was creating content for badges specifically so employed by
universities and their job was to actually write the description is to
create the badge images, but especially language work like writing the
description of let's say a course completion badge what it entailed, what
the achievement criteria are, what the skills the course demonstrated, that
sort of thing.

Dmitri Zagidulin: and I was a little shocked on this that there's enough
history and enough demand and enough complexity that this is what was
interesting to me. I was like, " this is really hard because these folks
straight up specialize in this." And so I found that interesting when
compared to sort of empowered assistance projects that are trying to bring
some of that expertise to a wider audience. as always I think we'll want to
talk about some of the interoperability initiatives and just in general we
want learner data to be portable between various wallets.

Dmitri Zagidulin: and just like we're talking about the data model of
verifiable credentials and education we want to talk about the sort of
natural complement which is APIs which is the protocols hey once we've got
all these credentials out into the hands of learners and earners and so on
the question then becomes how do blind parties consume those credentials?
How do they ask for those credentials? what sort of apps and what sort of
functionality do we have for helping people self-issue or curate their
credential collections into a story like what happens when you do a
curriculum VA a resume all that stuff falls under the API.

Dmitri Zagidulin: So we want to have some presentations on that and give
updates on Hey, so what are our friends the operating system vendors like
Google and Apple doing in terms of building in some of the verifiable
credential API into the browsers and into mobile OSS, right? what's the
sort of state-of-the-art in being able to request or issue credentials into
wallet applications? We want to continue the mobile driver's license
discussion. Simone added that we want to talk about Singapore migrating
from open search to OBB3. I believe MIT is slowly making that migration as
well.

Dmitri Zagidulin: which is funny because for example digital credentials
consortium of part of the team that I worked with is housed at MIT but just
because there's a dev and standards team there it's still harder to turn
the massive battleship that is a university as Ted mentioned in chat please
add yourself to the attendees list on the document And also please feel
free to raise your hand and…

Dmitri Zagidulin: and we'll add items of interest to the list. Phil says we
should ask Mark McConna of US Acro about the extent of adoption of CLRB2
for transcripts. Yeah, that's right. Let's talk about transcripts.

Phillip Long: It's not clear to me…

Phillip Long: how many institution actually we should ask him about the
status of it in two forms. One is ones that have actually done any pilot
work with it themselves. And secondly, if there are any that are routinely
using this as their default method for distingu trips to their students.
00:35:00

Dmitri Zagidulin: Excellent. Simony Infuse project. Yeah. that says Simone,
do you want to say a little more about that?

Dmitri Zagidulin: I'm actually not familiar with Infuse. huh.

Simone Ravaioli: Yeah, this is still an acros initiative.

Simone Ravaioli: I mean trying to catalyze a lot of the nonplayers in the
market around creating credentials data from unstructured files whether
it's syllabi and so on. It's been cooking in the back burner for a while.
But I do think it does touch the issue of credentials. It sparkle a little
bit of AI on top of it and maybe this group could be informed about what is
being publicly available because I do think that a lot of the tech that has
been talked about has to do with verifiable credentials and I mean
eventually might end up there.

Simone Ravaioli: So, it's just something that even Mark McConey,…

Simone Ravaioli: for example, as Phil mentioned, might give us an update on

Dmitri Zagidulin: Excellent.

Dmitri Zagidulin: From folks here on the call? what are some other topics
that you'd love to discuss?

Phillip Long: I've put in chat inferred skills from AI.

Phillip Long: That's one of the elements that is a part of infuse and I
don't know of any studies of any sort at this point that have actually
looked at the ability to determine the extent to which those inferred
skills are ones that have a high probability of the person they're being
attributed to performing.

Dmitri Zagidulin: So that's absolutely.

Dmitri Zagidulin: That would certainly be really interesting.

Dmitri Zagidulin: if such studies I suspect the field is moving too quickly
for the research community to keep up but that would be an amazing topic.
Phil, I think item 4 a captures what you're saying. but that's a great
point.

Phillip Long: I didn't see it. Or ah.

Dmitri Zagidulin: It's a very important topic because there's literally
multiple funded grants and…

Phillip Long: Yes. Yes.

Dmitri Zagidulin: initiatives and projects working on inferring skills from
text right now. All What else? raise your hand. mentions in chat where
would these inferred skills be derived from? Course job description.

Dmitri Zagidulin: Yes, exactly those two things and any other appropriate
sources, but course descriptions is what I see a lot of the project
focusing on.

Phillip Long: Yes. Ironically,…

Phillip Long: not LMS data.

Dmitri Zagidulin: That is ironic. Why is that? rather hold on. Is that
true? Aren't some of the LMS vendors also trying to participate in some of
these projects?

Phillip Long: Yes. …

Phillip Long: they're certainly sitting in on the infuse workg group.
whether they're interested in participating or not, I think remains to be
seen.

Dmitri Zagidulin: I see.

Dmitri Zagidulin: I see. what an interesting topic.

Phillip Long: But that opens a topic which is how do we get the data that
is actually where the performance of students in courses and their
assignments, their problem sets and the tests and all the rest of it are
managed.

Dmitri Zagidulin: Right. Right. Right.

Phillip Long: How does that provide a resource for skill inference that
might be more grounded? I mean there are many institutions I don't know if
many is the right word some institutions that are using either caliper or
open or xappi to generate events as a consequence of actions in the LMS
that go into an event log.

Phillip Long: And I know that's what they're doing at ASU as a research
tool, but I don't know how widely that's being actually examined or…
00:40:00

Phillip Long: whether those two methods are the only ways by which that
data is being aggregated from LMS's

Simone Ravaioli: Yeah, I'd be happy to pick this up and…

Simone Ravaioli: internally at the instructure find who the right person is
to speak about this and what instructure is doing. because I agree that any
effort that is happening out there to wishfully extract skills from a
course description you're still relying on the input not on the actual
projective.

Phillip Long: behavior.

Simone Ravaioli: So the moment we just continue to go back to it any
inference is going to be not very useful in the end.

Simone Ravaioli: And so clearly any system that collects that tracks
learning as it happens like the progressions and is deeper into picking up
any signals from completion or assessment data that feels like it's rich
with those kind of insights. but being from instructure I am also curious
about that and not an area that I touch directly but it is super
interesting and I know even in this past summer in July we started to
announce a few more things in that space so let me do some digging and…

Simone Ravaioli: I'll get back with someone that can come and share

Dmitri Zagidulin: Thank you.

Dmitri Zagidulin: by the goal.

Ildiko Mazar: And just another comment on this whole skills described in
course descriptions and…

Ildiko Mazar: jobs that require certain skills. it's one thing to indicate
the acquirable skills if one enrolls in a course or a program. But there's
another question which is probably the second level is the level of
proficiency in the acquired skills. So there can be a big difference
between two individuals having different proficiency levels of the same
skills and some jobs might need just average skills just demonstration of
some level of that skills and…

Ildiko Mazar: highly spec specific skills how do we square that circle?

Dmitri Zagidulin: Yeah, that's always a tough challenge for Go ahead.

Eric Shepherd: And just to let you know, the HR open is working on a
standard for the interchange of proficiency levels by skill.

Dmitri Zagidulin: Interesting. is there a link to that yet?

Phillip Long: Salad.

Dmitri Zagidulin: Excellent.

Eric Shepherd: I have the meeting schedule. Let me do some digging and I'll
try and put it in the chat.

Dmitri Zagidulin: So, we Fantastic.

Eric Shepherd: Okay. …

Dmitri Zagidulin: We may call on you in the future to present on it a bit.

Eric Shepherd:

Eric Shepherd: I'm not the guy, but I know the All right.

Dmitri Zagidulin: Okay. All right. Perfect. Yeah,…

Phillip Long: He's

Dmitri Zagidulin: as Phil mentions in chat, soon as you start moving the
inference stuff to let's say loadbearing use cases like medical students
always becomes perilous for example digital credentials consortium so the
team that I work with and Alex here on the call is doing

Dmitri Zagidulin: We got a small grant to work on again an open- source
badge creation tool assisted by LLM. So we're working on putting training a
small language model to help write let's say the description and…

Simone Ravaioli: That's a good question.

Dmitri Zagidulin: the criteria text for it. and the reason I bring it up is
it gets into the question that Ildico and others have mentioned of where's
the data set on which you train this and the answer is that data set is
difficult to come by. so we're reaching out to our members to various
stakeholders to say hey if you have interesting data sets of let's say
syllabus course descriptions and then the badges that were generated from
it we'd love to borrow that data to train this group model that people will
be able to use Phil.

Dmitri Zagidulin: Then yes.

Phillip Long: Yeah, if we're…

Phillip Long: if we're casting around for topics, the two that come to
mind, would that one that you and I are familiar with, but it is just
getting started, and that is creating a multi-skll credential from an
extension of the taking advantage of OBV3's ability to put in an array of
skills inside the credential. And we'll be looking at how it might be
possible to individually verify those otherwise rank them and performance
skills as performance capability.
00:45:00

Phillip Long: this is driven by employers who have expressed concern that
they are if individual skills are presented in OBV3s. it can be a lot of
things to look at from a given individual who might have 40 or 50 skills
that they have listed per application or per promotion request. and the
senior people doing that are saying we don't have time for how do we make
that something that's more digestible? so that's one thing.

Phillip Long: The other thing that comes to mind is the work going on in a
combination of collaboration with open and Jedex on the transmission of
structured data to states for the purposes of employment recording and that
uses OBV3s and similar credentials for which there has been two pilots.
Eric, I know that you've hosted one of those on the CCC web webinar series.

Phillip Long: So that might be one that's worth bringing to this audience
as well.

Dmitri Zagidulin: Thanks, Phil.

Dmitri Zagidulin: Yeah, those all sound like excellent topics. What else? I
do wonder if the time has come to start talking about terms of use for
prepare for credentials that issuers or even better yet the holders the
students themselves might be able to say yes I'm sharing this data with you
because you requested but here are the restrictions on I would like for you
to not resell it to third parties.

Dmitri Zagidulin: is that sort of thing.

Phillip Long: I am aware that A4L has come up with…

Phillip Long: what sounds like a terms of use document to that would
accompany a credential. we should probably ask Alex Jackal to present on
that or whom he would suggest present on that.

Dmitri Zagidulin: All right, sounds good.

Simone Ravaioli: I don't know.

Dmitri Zagidulin: Yeah, please feel free to reach out to him.

Phillip Long: I can reach I'll take that one on. What else? I don't know if
this is in scope or not but there is the learning and work encyclopedia
folks that are building a set of resources the context of a learning and

Phillip Long: work library that might be worth bringing to the attention of
this group…

Phillip Long: since it is a reasonably up-to-date compilation of papers and
project reports and things like that that do intersect with verifiable
credentials. Holly Zaneville is the person…

Dmitri Zagidulin: That does sound interesting.

Dmitri Zagidulin: Any fantastic let me capture…

Phillip Long: who leads that. I'll take that one on as well since I know
Holly.

Dmitri Zagidulin: what Eric posted and Hildo. Yeah, that would be
interesting to hear how works with skill tagging and if there's any data
sets. go ahead. Go ahead.

Ildiko Mazar: As a matter of fact, one of the events that is in the
announcements and reminders, the one attempu conference next week, we are
going to have a session…
00:50:00

Phillip Long: I

Ildiko Mazar: where we are trying to come to agreements about a set of open
badge extensions that would be aligned with these favorable properties.

Simone Ravaioli: That's

Ildiko Mazar: So if you want us to report after that that would be probably
very interesting because as you all know the European audience is very
eager to u be compliant with the European Council recommendation of micro
credentials that does have a very strict expectation of …

Ildiko Mazar: how learning outcomes are described and those refer to
skills. So that would be interesting.

Dmitri Zagidulin: Yeah, thank you.

Dmitri Zagidulin: I'd love to hear about it. That sounds really
interesting. Eric, is there a way to open up the access to that Google doc
or…

Eric Shepherd: I'm trying to do that right now.

Dmitri Zagidulin: would you need to Okay,…

Eric Shepherd: They should all be open, but I'll work on it.

Dmitri Zagidulin: What else has caught your eye? topics and questions as
you were encountering credentials over the summer on the CCG.

Dmitri Zagidulin: There was a poll that went out ongoing for several weeks
and…

Simone Ravaioli: Oops.

Dmitri Zagidulin: I think this is the last call for it that Manu sent out
topics for standardization for this upcoming verifiable credential working
group. And so render method is going to be one of those topics or at least
I highly suspect it's some of the other topics on that poll that might be
interesting to this group are offline credentials basically representing
and compressing credentials into QR codes for use cases.

Dmitri Zagidulin: And the other one is verifiable credentials over
Bluetooth a protocol being able to exchange credentials again in a offline
friendly secure way of between mobile devices.

Simone Ravaioli: Good day.

Dmitri Zagidulin: So, Eric mentioned link should be open. Let's take a look.

Dmitri Zagidulin: No, it is not requested access.

Eric Shepherd: Can I email it to you and…

Eric Shepherd: then you can distribute it? It's meant to be open.

Dmitri Zagidulin: Yeah. Of course.

Eric Shepherd: Okay.

Dmitri Zagidulin: Of course. Yeah, please.

Eric Shepherd: Thank you.

Dmitri Zagidulin: All right. last call for items. And of course, this is
not the last opportunity for input. If you come up with other topics you'd
like to present on or for the group to discuss, please email the mailing
list. "Hey, let's start a conversation about

Dmitri Zagidulin: And if there's no more items, let's adjourn.

Eric Shepherd: Thanks. Right.

Dmitri Zagidulin: Everybody needs time before the next call to hop on.
Thanks Great to see everyone and talk to you in the coming weeks.

Ildiko Mazar: Have a nice day and evening to those in Europe. Bye.
Meeting ended after 00:56:34 👋

*This editable transcript was computer generated and might contain errors.
People can also change the text after it was created.*