Universal Skills Authentication

Deliverables and Definition of Success

Deliverables would be sufficient to provide an MVP for universal skills authentication using Atala PRISM:

1. skills framework: flexible data structure for describing individual skills in a graph-based format (use of a graph allows connections between skills), which can be stored on any platform off-chain (usual approach) or on-chain (a 'personal skills graph') -- this would be based on existing LearnerShape approaches to skills (see below)
2. individual authentication: protocol and REST API for self-sovereign authentication of a personal skills graph to an individual, using Atala PRISM -- this would aim to be a simple specification (to be enhanced later) drawing upon and ideally compatible with existing work on authentication of credentials (notably the W3C Credentials Community Group (CCG) specifications: https://w3c-ccg.github.io/community/work_items.html#completed-specifications))
3. skills authentication: protocol and REST API for a third party (e.g. a university, course provider, examination institute or another individual) to certify that an individual has a skill on their personal skills graph, using Atala PRISM -- like #2 this would aim to draw on existing work like that of W3C CCG
4. application for skills entry, certification and verification: simple web application for accessing the above, including (a) building and certifying a personal skills graph, (b) third-party certification of a skill in the graph and (c) viewing / confirming a personal skills graph.

Together these components would allow an individual to create and share an authenticated personal skills graph, with any number of those skills certified by one or more third parties. Authentication of skills to an individual using DID based on Atala PRISM (#2 above) would allow third parties to provide certifications (#3 above) with confidence that they are certifying the right person (and that such credentials could not be improperly reused).

We would implement these deliverables primarily in Python with a PostgreSQL database and access via a REST API, like the rest of the LearnerShape open source infrastructure (see below). For integration with Atala PRISM and other Cardano infrastructure, we have agreed to partner with Gimbalabs (see Team section below). The specific requirements for such integration are not yet clear, because the Atala PRISM API and SDK have not yet been released.

We would define success in this project as:

3 months

• fully-articulated development plan for all deliverables
• substantial progress on coding of deliverables

6 months

• completion of deliverables and public launch, with code available as open source on LearnerShape Github repository (and within Cardano infrastructure as appropriate)
• plan for go-to-market through PeopleCert, Project Catalyst teams and other partners (see Go-To-Market Plan section below)

12 months

• working on further roadmap to roll out universal skills authentication in partnership with Project Catalyst / IOHK / IOG, PeopleCert and other partners.

Because the Atala PRISM API and SDK have not yet been publicly released, this proposal is for a proof-of-concept implementation that would be further developed as Atala PRISM advances. We would hope to better understand the technical details of Atala PRISM in the course of this project (ideally obtaining details of the API and SDK), though interactions with Project Catalyst and broader IOHK/IOG teams, to advance integration with Atala PRISM as far as possible at this stage.

Future development beyond this project could include (among others):

• fully functional web and mobile apps to access the skills authentication protocol
• allow certifying bodies to revoke a certification
• allow individuals to present only a subset of their personal skills graph
• integrate LearnerShape machine learning functionality for skills comparison into the skills graph
• consider incentivizing use of the skills authentication protocol with ADA and/or a bespoke token
• integrate (and support interoperability) with other authentication technologies (e.g. new 'personal ID' feature to be added to iOS 15).

Skills Authentication and Potential for Driving DID Mass-Scale Adoption

The accurate communication of skills is a difficult challenge. Whether the skills taught by a university course or the skills of a job applicant, verification is often incomplete, laborious and susceptible to significant errors or outright fraud.

There are various ways that skills can be learned and confirmed, e.g. (a) diplomas and certificates from universities and other educational institutions, (b) certificates from professional bodies, certification organizations and examination institutes, (c) references from colleagues, mentors or supervisors, and (d) evidence or demonstration from the individual with the skill.

Modernizing skill certification with Atala PRISM would create a universal system in which

• any individual can claim a skill or qualification
• another entity, such as a university, can certify they have that skill
• the certification can be verified by another entity, such as a prospective employer.

The universality of such a system relies on a common framework for skills. There is extensive prior development of skill frameworks, including options for the entire economy (e.g. O*NET and ESCO) and specialized options for specific industries and fields. However, the lack of any generally-accepted skills framework and the periodic announcement of new frameworks produces significant confusion. LearnerShape has addressed this challenge by using AI to enable organizations to work with any skills framework (see https://www.learnershape.com/blog/going-beyond-skills-taxonomies-with-AI). LearnerShape's approach allows a skills certification in one framework to be compared with a skills certification in any other framework, allowing the overall system to be universally applied.

The potential for Atala PRISM and DID generally to address the legitimacy of skills and credentials is already being explored, including through the Ethiopia project (which is a major inspiration for this proposal) as well as other proposals in response to this challenge (we are cooperating with some of these – see Team section below). Expanding authentication using Atala PRISM to the huge market for individual and workforce skills would provide a powerful driver to mass-scale adoption of Atala PRISM. By 2025, the global education market is expected to reach ~$7.3T including ~$404B for educational technology (https://www.holoniq.com/notes/global-education-technology-market-to-reach-404b-by-2025/), and digital skills management will be a key enabler of this ongoing growth.

LearnerShape and Open Source Learning Infrastructure

LearnerShape (https://www.learnershape.com/) is building the world's first AI-driven, open source learning infrastructure – a set of microservices for delivering any education application. Our open source components are in the lsgraph repository (https://github.com/LearnerShape/lsgraph), and an explanation of our services is in our e-book (https://mailchi.mp/learnershape/ebook-april-2021-flexible-ai-based-open-source-learning-infrastructure).

A key challenge in recommending a future career or the best learning resources is understanding the current skills of the individual. An area of focus for LearnerShape has been how to work with the different concepts, terms and skills frameworks, taxonomies and graphs currently in use. Adding skills certification and authentication is a key piece of this process, and we hope to work with Cardano Catalyst and Atala PRISM to do so. Our previous work on flexible skills frameworks (see e.g. https://www.learnershape.com/blog)) will give us a major head start on the proposed project.

We provide a YouTube link below to a LearnTech Meetup event at which Maury Shenk speaks about LearnerShape's approach to skills at 33:40 to 41:04.

Team

Our key team members are:

Maury Shenk (proposer, https://www.linkedin.com/in/mauryshenk/)) is CEO and co-founder of LearnerShape, director at PeopleCert (a global testing and certification company) and experienced technology lawyer. He has worked on digital authentication issues for more than 25 years. He would lead the proposed project.

Jonathan Street (https://www.linkedin.com/in/jonathanstreet/) as Head of Data Science at LearnerShape has been applying advanced machine learning models to understand the relationships between skills. He would lead technical implementation of the proposed project.

Sean Miller (https://www.linkedin.com/in/seanmiller1066/) is an experienced full-stack developer who plays a key role in delivering LearnerShape applications. He would assist with interface elements of the proposed project.

Lorenzo Zorzi (https://www.linkedin.com/in/lorenzo-m-zorzi-chiarioni/) is an advisor to LearnerShape with strong entrepreneurial experience and a background in financial services. He would assist with designing the user proposition for the proposed project.

We have also agreed to cooperate with Gimbalabs for integration with Atala PRISM and the Cardano ecosystem more broadly, including Haskell programming and general experience with the Cardano environment. This cooperation has been agreed through discussions with Randall Harmon (https://cardano.ideascale.com/a/pmd/3084778-48088?) and other members of the Gimbalabs team.

Budget and Funding

We propose a budget of USD 48,000. This is primarily for the following development costs:

• design of flexible skills framework, including data structures and file formats (deliverable #1) -- USD 10,000
• individual authentication protocol based upon Atala PRISM (deliverable #2) -- USD 7500
• skills authentication protocol based upon Atala PRISM (deliverable #3) -- USD 7500
• web application for accessing the above (deliverable #4) -- USD 18,000.

In addition, we have budgeted USD 5000 for user proposition design and project management. These amounts include our costs of cloud infrastructure (primarily compute and database on Amazon Web Services) and other overhead.

The proposed budget is below the cost for delivering the project on a commercial basis, but we are committed to co-investment of time and resources to work with Cardano on this project.

Later stages of the project could be funded by Project Catalyst or other sources (e.g. PeopleCert, angel/VC investors or grants). LearnerShape is an early-stage company. Our initial development has been funded by Innovate UK (the UK's government innovation agency) and founder/friends equity. We are currently operating on a per-project bootstrap basis, and anticipate raising seed funding once we have sufficient commercial traction.

Go-To-Market Plan

Once a universal skills authentication protocol is developed, we would actively promote its use.

Our key partner for going to market is PeopleCert (https://peoplecert.org/), a global provider of certification services where Maury Shenk is on the board of directors. PeopleCert has been cooperating with LearnerShape since 2019, and has specifically agreed to support LearnerShape in its cooperation with Project Catalyst. PeopleCert has recently substantially strengthened its market position through acquisition of AXELOS, a joint venture between Capita and the UK Cabinet Office (see https://www.capita.com/news/agreement-to-sell-stake-axelos).

There is also a significant opportunity for cooperation between our project and other Project Catalyst projects. During the F6 proposal refinement process, we have already discussed cooperation opportunities with proposers of various related F6 projects including:

• Blockademia Verification System (https://cardano.ideascale.com/a/dtd/Blockademia-Verification-System/366827-48088) – dApp for certification of credentials; strong potential synergies with our project as a third-party certifier (see deliverable #3 above)
• Optimizing student agency with DIDs (https://cardano.ideascale.com/a/dtd/Optimizing-student-agency-with-DIDs/369878-48088) – certification of student achievements and progress; potential use case for our proposed service
• Education verification with DID (https://cardano.ideascale.com/a/dtd/Education-verification-with-DID/368554-48088) – certification of university credentials, starting in Malaysia; potential use case for our proposed service.

The proposers of all of these projects have confirmed initial interest in cooperation. We would pursue these collaborations once our project has advanced towards public launch, and we expect that numerous other cooperation opportunities will arise in the very dynamic Project Catalyst environment.