How to Certify AI Skills with Microcredentials: The Universidad Ean Case

In 2026 almost anyone can show a certificate from an artificial intelligence course. Almost none of them prove anything. And this field carries an irony of its own: the very technology being learned can fabricate the evidence of having learned it. Universidad Ean approached the problem from the right question, which is not how to certify an AI course but which AI skill is worth recognizing.

Direct answer: certifying artificial intelligence skills takes three things an attendance certificate does not provide. Naming the applied skill instead of the course or the tool, assessing evidence that the person built something, and issuing the result as a digital credential that can be verified independently. Universidad Ean, in Bogotá, Colombia, implemented this with POK Proof of Knowledge: credentials that recognize AI prototyping and solutions applied to real problems, issued to students, faculty and staff, verifiable in one click from a public page, and a large share of them backed by an NFT on blockchain. A single program certifies participants from Colombia, Mexico and Ecuador with the same credential.

Key points

• Tools expire, skills do not: a certificate naming a tool loses meaning as soon as that tool ships a new version. One naming the skill, prototyping solutions with AI, still says something two years later.
• AI is the one field where AI can forge the evidence: résumés, portfolios and even code are generated in minutes. A credential issued and verified by the institution is the proof that cannot be fabricated.
• The skill is demonstrated by building: Ean's AI credentials are issued on prototyping programs and on real company challenges, not on classroom hours.
• Certify the people who teach, not only the people who learn: the university's AI programs include students, faculty and staff, and everyone receives a credential.
• One credential, three countries: an international consulting program brings together participants from Colombia, Mexico and Ecuador, and all of them get the same verifiable credential wherever they are.

The problem is not certifying AI, it is deciding what gets certified

When an institution decides to move into artificial intelligence training, the conversation almost always starts with the catalog: which courses to open, with which tools, how many hours. Certification is left for the end, as the paperwork that closes the process. That is where the mistake slips in, because the question of what the credential says is really the question of what the program teaches.

A certificate reading "Generative AI course, 20 hours" has an expiration problem. It names a technology that will work differently a year from now, and it says nothing about what the person can do. The employer receiving it is left exactly where they started: they still have to find out on their own whether the candidate can apply anything.

A credential reading "AI Prototyping and Solutions" describes something else. It names a capability rather than a tool, and that capability survives changes of model, vendor and interface. It is the difference between certifying exposure and certifying competence, and it is a curriculum design decision before it is a technology one.

The urgency is not theoretical. The World Economic Forum's Future of Jobs Report 2025, built on a survey of more than a thousand employers representing over 14 million workers, projects that 39% of the key skills in the labour market will change or become obsolete by 2030, and it ranks AI and big data first among the fastest-growing skills. That is the most uncomfortable pairing possible for whoever issues the certificate: the fastest-growing area is also the fastest to go out of date.

	Certificate from an AI course	Verifiable skill credential
What it names	The tool or the course	The applied skill
When the model shifts	It loses currency	It stays legible
What it evidences	Attendance and hours	An assessed deliverable
How it is checked	Someone at the institution answers mail	A public verification page, in one click
Who owns it	The downloaded file	The person, in their digital wallet
Adds up to more	No	Stacks toward larger recognitions

"The conversation today is no longer about issuing certificates, but about which competencies are worth recognizing."

Juan Manuel Medina Barrera, Curriculum Development and Assessment Center Coordinator, Universidad Ean

That job title is worth pausing on, because it says something about how the university organized the problem: curriculum development and assessment center inside the same role. Defining what gets recognized and defining how it gets assessed are two sides of the same work, and when they live in separate departments the usual result is a course catalog with certificates nobody designed.

The irony of the field: AI can fabricate its own evidence

There is an argument for verifiable credentials that lands harder in artificial intelligence than in any other discipline.

In 2026 a candidate can use AI to generate a flawless résumé, a convincing LinkedIn profile, a portfolio of projects that never existed, and even snippets of code that look like their own. Every traditional signal a recruiter used to assess technical skill became fabricable, and it is especially fabricable in AI itself, where the tool the person supposedly masters is the same one producing the fake proof. We covered this problem in detail for the AI-written résumé, and in AI training it becomes almost circular.

What cannot be generated is a credential issued by an identified institution, with a public page where anyone confirms who issued it, to whom, and for which skill. Not because the file is hard to copy, but because verification does not consult the file: it consults the issuer. Anyone can make an image that says "Universidad Ean". Nobody can make Universidad Ean stand behind it.

So in the field where evidence is easiest to fake, a verifiable credential is not a presentation detail. It is the only signal left standing.

Five territories, one grammar of recognition

Universidad Ean, in Bogotá, holder of Colombia's High Quality Institutional Accreditation and with more than five decades of history, did not digitize a certificate: it designed a system. Its credentials recognize concrete competencies in the territories where the university wants to set the agenda, and artificial intelligence is first on that list.

• Applied artificial intelligence. AI prototyping and solution programs, with students, faculty and staff taking part.
• Entrepreneurship. Venture creation and acceleration tracks, in line with the DNA of a university known as the home of sustainable entrepreneurship in Colombia.
• Sustainability and innovation. Consulting applied to real company challenges.
• Creative and cultural industries. Training alongside players from the Ibero-American cultural ecosystem.
• International consulting. Collaborative challenges that bring together participants from different universities across the region.

The real credentials coming out of this have names describing what the person did: "AI Prototyping and Solutions", "Consulting on Innovation for Corporate Sustainability", the "Impacto Maker" badge. None of them is the name of a piece of software.

What makes AI work as a certifiable territory is that the programs rest on output: prototyping, solving a company challenge, delivering something. An applied skill is assessed on what the person built, and that gives the credential something it can name precisely. A theoretical course about AI, by contrast, leaves the institution with only one thing to certify, which is attendance.

Certify the people who teach, not only the people who learn

Most university AI literacy programs are designed for students. At Ean, the AI programs also include faculty and staff, and everyone receives a credential.

Faculty AI competence is also already described precisely outside any single university. In 2024 UNESCO published two reference frameworks, one of AI competencies for teachers, with 15 competencies across five dimensions, and one of AI competencies for students, with 12 across four. So there is an international vocabulary for naming what is being recognized, which is exactly what a certificate reading "AI course, 20 hours" does not have. An institution that wants its AI credential to mean something beyond its own campus has a map to align with.

The reason goes well past internal consistency. When the teaching body goes through the same program it is about to teach, and receives the same credential it is about to issue, two things fall into place on their own. The first is adoption: nobody has to sell a professor on the value of a digital credential he already has published on his profile. The second is judgment: whoever assessed the skill from the learner's side knows what is reasonable to ask before recognizing it.

In artificial intelligence this weighs more than in any other area, because it is a field where many students arrive with more hands-on practice than the institution has. A teaching body that never went through the program ends up assessing something its own students handle better.

One credential, three countries

One of Ean's programs brings together participants from universities in three countries of the region: Colombia, Mexico and Ecuador. All of them receive the same verifiable credential, regardless of where they are enrolled.

It reads like a logistics detail and it settles considerably more than that. The traditional way to certify an academic experience shared across institutions in different countries is one document per institution, each with its own format, signature and validation circuit, and none of them legible outside its own system. An employer in Quito holding a certificate issued in Bogotá has no practical way to check it, so as a rule they do not try.

A verifiable digital credential solves that without mutual recognition agreements: verification is the same operation in all three countries, it happens on the credential's public page, and it does not depend on the verifier knowing the issuing institution or speaking its administrative language. The credential crosses the border because it needs nobody on the other side to translate it.

What makes an Ean credential verifiable

The real shift from the signed PDF certificates the university used before is about who can check the credential, and when.

One-click verification. The employer opens the credential link and confirms who issued it, to whom and why. No calls, no emails, no waiting. Before POK every validation required a human: someone at the university had to answer each request, one at a time, on turnaround times nobody is willing to wait for today. This is how employer-side credential verification works now.

Owned by the person. The credential lives in the student's digital wallet and follows them through their whole professional career. The PDF did the opposite: it arrived by email, got downloaded, and its useful life ended there, filed away in a folder.

Public, tamper-proof record. A large share of the university's credentials is also issued with NFT backing on blockchain, a record anyone can consult and nobody can alter after issuance, including the institution itself.

Open standards. POK issues under Open Badge 3.0, the standard that defines how a digital credential is described and signed so it can be checked without depending on the platform that issued it. That is what keeps the credential from being locked to a vendor, and the conformance can be audited independently of us: POK is listed in the 1EdTech certified product registry, with active certification for Open Badges 3.0 and Comprehensive Learner Record.

From the classroom to the credential

The full journey inside POK has four moments, and the third one tends to surprise institutions that are just getting started.

1. It is earned. The program or learning experience.
2. It is issued. The digital credential goes out in minutes, and full cohorts are certified in a single operation, with no friction for internal teams. Before, every new cohort meant repeating a manual process end to end.
3. It is shared. The graduate posts it on LinkedIn, adds it to a résumé, or drops it into a portfolio.
4. It stacks. The credential adds up toward larger recognitions.

The third step is the one that returns value to the institution without the institution doing anything else. Every credential shared on LinkedIn turns a private achievement into a professional showcase, and the university's brand travels with it, verified. A PDF in a folder does none of that.

"With POK we went from handing out documents to offering verifiable digital evidence."

Universidad Ean

What changed

The most visible change is not in the platform, it is in what people do with their credentials. Ean students open them, download them and share them on their professional networks, which is exactly what a PDF certificate never prompted.

• Recognition that becomes visible. Every credential shared on LinkedIn turns an achievement into a public, checkable signal, right where candidates are sourced.
• Zero friction for the graduate. No paperwork, no emailing the university for proof of completion.
• Institutional reputation. The Ean brand travels with every verified credential, on every graduate's profile.
• Internal culture. Faculty and staff get certified too and adopt the model, so recognition of learning stops being something the institution hands out and becomes a practice that runs through it.

Worth pinning down the scope: these are process and behavior changes, measurable in adoption and friction, and not yet an employability metric. The university also did not pick the platform looking for a software vendor, but for a partner to support a change of model.

"POK allowed us to move from issuing certificates to building a genuine ecosystem for recognizing competency-based learning. It is a reliable, flexible platform, aligned with what the education of the future needs."

Antonio Alonso González, Vice Rector for Academic Innovation, Universidad Ean

What comes next: from the standalone certificate to skill pathways

Ean is already working on the next stage, the one that turns a credential catalog into a system.

• Stackable credentials. Microcredentials that add up to each other and build complete learning pathways. This is the step where recognition of learning starts having academic consequences, and where institutional design weighs more than technology: credit recognition rules are best settled before scaling, because revising them once thousands of credentials are out forces a decision about every one already issued.
• New focus areas. Artificial intelligence, sustainability and entrepreneurship as the axes of the catalog.
• Recognition of prior learning. Certifying what was learned outside the classroom, at work and in professional life. In AI this is where it is needed most, because a huge share of real skill was picked up by using the tools, not by taking courses.
• The whole trajectory. From the short workshop to the doctorate, with credentials accompanying every stage.

How to certify AI skills at an institution

The order worth following is replicable, and it is almost the reverse of what usually happens.

1. Name the skill, not the tool. Write the credential name before building the program. If the name does not survive the next model version, it is not a skill yet.
2. Design the program around a deliverable. A prototype, a company challenge, an applied solution. Without output there is nothing to assess beyond attendance.
3. Define who assesses and against what criteria. Before the first issuance, not after the first complaint.
4. Put faculty and internal teams in the first cohort. It solves adoption and calibrates the assessment criteria.
5. Issue verifiable from day one. A public verification page and an open standard. Migrating later means reissuing everything already issued.
6. Design stackability up front. It is far easier to define how three credentials add up while none have been issued yet.

For the full starting point beyond the AI case, the step-by-step guide to implementing microcredentials at a university covers the process in detail.

The university's programs and credentials can be explored at universidadean.edu.co.

Frequently asked questions

How do you certify artificial intelligence skills?

By naming the applied skill instead of the course or the tool, assessing a concrete deliverable that demonstrates it, and issuing the result as a verifiable digital credential with a public verification page. A certificate naming a tool loses value when the tool ships a new version; a credential naming the capability, for example prototyping solutions with AI, is still legible years later. Universidad Ean follows this model with credentials issued through POK Proof of Knowledge.

Is a certificate from an artificial intelligence course worth anything?

It works as proof that the person took the course, and not much more. The problem is what it names: a certificate reading "Generative AI course, 20 hours" describes a technology that will work differently before long, and says nothing about what the person can do. The World Economic Forum projects that 39% of the key skills in the labour market will change or become obsolete by 2030, and ranks artificial intelligence first for growth. A credential that names the applied skill and is verified on a public page survives that shift; an attendance certificate for a course about a tool does not.

What is an AI competency framework?

It is a document defining what a person needs to be able to do in order to use artificial intelligence competently, broken down into concrete competencies. UNESCO published two reference frameworks in 2024: the one for teachers defines 15 competencies across five dimensions and the one for students 12 across four, among them a human-centred mindset, the ethics of AI and AI system design. They let an institution name precisely what each credential recognizes, instead of certifying attendance at a course.

Why does a verifiable credential matter more in artificial intelligence than in other fields?

Because it is the one field where the technology being learned can fabricate the evidence of having learned it. AI generates a flawless résumé, a portfolio of nonexistent projects or someone else's code in minutes. Every traditional signal of technical skill became forgeable. A verifiable credential is not checked by looking at the file but by consulting the issuer, so it cannot be fabricated without the institution.

Which AI credentials does Universidad Ean issue?

The university issues credentials on programs for AI prototyping and solutions applied to real problems, with students, faculty and staff taking part. Real credentials include "AI Prototyping and Solutions", "Consulting on Innovation for Corporate Sustainability" and the "Impacto Maker" badge. Artificial intelligence is one of the five territories in its catalog, alongside entrepreneurship, sustainability and innovation, creative and cultural industries, and international consulting.

Can the same credential be issued to students from several countries?

Yes. One of Universidad Ean's programs brings together participants from universities in Colombia, Mexico and Ecuador, and all of them receive the same verifiable credential. Verification is the same operation in all three countries, it happens on the credential's public page, and it requires no mutual recognition agreements and no prior familiarity with the issuing institution.

What is the difference between a PDF certificate and a verifiable digital credential?

A PDF depends on a person to be validated: someone at the institution answers each request, one at a time. A verifiable digital credential is checked in one click from a public page, with no intermediaries. The PDF also arrives by email, gets downloaded and ends its useful life there, filed in a folder, while the credential lives in the person's digital wallet, is shared on LinkedIn and on résumés, and can add up toward larger recognitions. A large share of Ean's credentials is also issued with NFT backing on blockchain, a public record nobody can alter after issuance.

Is it worth certifying faculty in AI skills too?

Yes, and in artificial intelligence more than in other areas. When the teaching body goes through the same program it is about to teach and receives the same credential it is about to issue, internal adoption stops needing persuasion and the assessment criteria get calibrated by someone who experienced the program from the learner's side. Universidad Ean's AI programs include students, faculty and staff.

Related reading

• Can you trust an AI-written résumé? Why verifiable credentials are hiring's new proof of truth
• How to implement microcredentials at a university, step by step
• The real value of digital credentials: from commodity to trust infrastructure
• Open Badge 3.0: the complete guide to digital credentials

To see how POK works with institutions across Latin America, take a look at plans and pricing or book a demo.

Last updated: August 21, 2026.