Educators are deciding which AI skills to teach. How do they know?

Accreditors are deciding which programmes meet the standard.

Employers are deciding which candidates are genuinely ready.

And individuals ­­­­­­– millions of them ­­– are considering career decisions and self-education trying to guess what is going to make them and keep them employable.

Everyone in this chain is trying to make an informed decision. Who has the information they actually need to make it well?

Because the question underneath all of it, the one that keeps getting skipped, is not which AI skills matter. It's how you know whether someone actually has them.

Something significant is happening in higher education right now that brings this into sharp focus.

Purdue University has become the first US institution to require every undergraduate ­­– regardless of major ­­– to demonstrate AI competency before they can graduate.¹

Ohio State has gone further, committing that every student from the class of 2029 will leave not just knowing about AI, but genuinely fluent in applying it within their field of study.²

These are serious institutions making serious public commitments. This isn't an elective. It isn't a pilot. It's a graduation requirement.

And it's coming. To Europe. To Asia. To every university that wants its graduates to be employable in an AI-shaped economy.

But when a university says its graduates are AI fluent ­­– how do you know? Not how do you hope. Not how do you assume based on the curriculum design document. How do you actually know?

Because the policy is not the evidence. The learning outcome written into the programme specification is not evidence of competence, that's evidence of an input. And the course completion data sitting in the LMS is definitely not the evidence.

As Alex Kotran, CEO of the AI Education Project, put it when Purdue made its announcement: "This is promising, but the proof will be in the pudding: Are they building the systems for professional development and learning? The policy is just step one."³

Step one.

Universities are racing to make the commitment. The infrastructure to verify it ­­– rigorous, contextualised, practice-based assessment that actually distinguishes genuine fluency from module completion ­­– is nowhere to be seen. Yet.

The university made a commitment. The accreditor approved it. The individual invested time and money trusting it would be worthwhile. None of that answers whether the credential they're holding justifies what they paid for it today; or whether it will be worth anything tomorrow?

That matters for employers. It matters for professional bodies whose members are emerging from these programmes. And it matters enormously for the organisations whose quality assurance stamp sits on top of the claim.

Are the accreditors who put their stamp on these programmes ready to answer that question?

Part 1 of 3. More on that soon.

¹ Purdue University Board of Trustees, December 2025 

² Ohio State University AI Fluency Initiative, June 2025 

³ Alex Kotran, CEO, AI Education Project, quoted in The 74, February 2026