The University After AI

When knowledge is no longer scarce

A structure built on scarcity

The university was never just a place. It was a structure. Built on scarcity—of books, of expertise, of access. To study meant to enter a space where knowledge was concentrated, curated and transmitted. Lecture halls, libraries, journals—these were not just facilities. They were filters. They defined what counted as knowledge and who was allowed to engage with it.

There was also an implicit contract. If you did the work—if you read, struggled, wrote, refined— you would earn the status of someone who knows.

That structure is now dissolving. The university was built to organise access to knowledge. It must now redefine how knowledge is understood.

Not slowly. Not gradually. But asymmetrically—faster at the edges, uneven at the core. Artificial intelligence does not simply add a new layer to the university. It breaks the historical link between effort and intellectual outcome.

Knowledge is no longer scarce. And with that, the logic of the institution begins to shift.

From access to abundance

For centuries, universities operated as gatekeepers of information. Access required presence. Interpretation required training. Synthesis required time.

Now, much of that can be generated on demand.

Summaries appear instantly. Arguments can be structured in seconds. Entire essays can be drafted before the first real question has been fully understood. What once required effort is now increasingly frictionless.

But friction was never a flaw. It was the mechanism of learning.

“The challenge for universities is no longer to provide access to information, but to cultivate the capacity for critical judgment in an age of information abundance.”
Andreas Schleicher, Director for Education and Skills, OECD

The implication is subtle, but profound. If information is abundant, then the value of education shifts.

Not towards more information, but towards something else entirely: judgment.

An institution under pressure

This is where the tension becomes visible.

The university has always had multiple roles. It transmits knowledge. It produces knowledge. But it also forms judgment—often implicitly, sometimes unintentionally, but always structurally.

Artificial intelligence disrupts the first two.

It accelerates access.
It augments production.

But it does not replace responsibility. And that is where the institution is now under pressure.

Europe in three responses

Across Europe, universities are responding—but not in the same way. Some move early. Some hesitate. Some resist.

At the University of Helsinki, artificial intelligence is treated not as a threat, but as a public capability. The widely adopted “Elements of AI” programme was designed not just for students, but for citizens. The premise is simple: if AI is reshaping society, then understanding it cannot remain confined to specialists.

“AI literacy is the new civic skill. Our goal is to demystify AI so that it doesn’t just happen to us, but is shaped by us.”
Teemu Roos, Professor Computer Science, University of Helsinki

Here, AI is positioned as a layer of collective knowledge—something to be understood, not merely used.

Elsewhere, the response is more exploratory.

At TU Delft, the question is not whether AI will be used, but how far it should be allowed to shape engineering processes. Students are already integrating AI into design, modelling and analysis. But this raises a deeper concern: if systems begin to generate solutions, what remains the role of the engineer?

“AI is a ‘power tool’ for the mind, but a tool is only as good as the person wielding it. We must teach students not just how to use AI, but how to remain the ultimate authority over the systems they create.”
Tim van der Hagen, Rector Magnificus, TU Delft

The emphasis shifts from capability to control. From execution to oversight. And then there are institutions where the first response was to draw a line.

At Sciences Po in Paris, generative AI was initially banned outright. The concern was not technological, but academic: if students could outsource writing and analysis, what remained of intellectual development?

“Using ChatGPT for oral or written work is strictly prohibited, as it raises questions of integrity and plagiarism. The goal of education is to ensure that students are able to develop their own thinking.”
Sergei Guriev, Provost, Sciences Po Paris

That position has since evolved.

Not into full acceptance, but into a hybrid model—where use is permitted under strict transparency. Students must disclose how AI was used, which prompts were given and where machine assistance ends.

The shift is telling. The first instinct was control. The second is adaptation.

The luxury of struggle

This is not a European weakness. It is a European instinct. To regulate. To preserve. To question before accelerating.

But the deeper question is not how to control AI. It is what remains of learning when efficiency becomes dominant.

Education has always depended on friction. The slow search for sources. The rewriting of sentences. The structuring of thought over time.

The long hours in archives. The iterative feedback on a thesis. The tension of a disputatio, where ideas are tested in argument, not generated on command.

These are not inefficiencies. They are the process through which understanding emerges.

Artificial intelligence removes much of that friction.

Struggle was not a barrier to learning. It was the environment in which learning took place.

The sweat on archival documents is replaced by a prompt. The hesitation before a paragraph becomes a suggestion in seconds.

And in doing so, it introduces a paradox: The more efficient learning becomes, the less certain it is that learning is actually taking place.

Efficiency, in this context, is not neutral. It can become the enemy of education.

What is at stake

What, then, is at stake? Not just curriculum. Not just assessment. But the definition of knowledge itself.

If knowledge can be generated, is it still something to be acquired? If answers are always available, what is the value of asking questions? If writing can be delegated, where does thinking reside?

These are not rhetorical questions. They are institutional ones.

Between performance and understanding

The answer is not to reject artificial intelligence. That would be both unrealistic and counterproductive.

AI is already embedded—in tools, in workflows, in expectations. Students will use it. Researchers will rely on it. Institutions will integrate it.

The question is not whether. But how. And more importantly: Where to draw the line.

A functional approach is not enough.

Yes, AI can improve efficiency.
Yes, it can support understanding.
Yes, it can expand access.

But it can also simulate competence.

A well-structured answer is not necessarily a well-understood one. A coherent essay is not necessarily the result of coherent thinking.

This creates a new challenge for universities: To distinguish between performance and understanding. Between output and insight.

From knowledge to judgment

This is where the role of the university becomes clearer again. Not as a provider of information. But as a validator of judgment.

Students must still learn how to define problems, assess sources, recognise assumptions and take responsibility for conclusions.

Not because AI cannot assist in these tasks. But because AI cannot be accountable for them.

A European direction

In this sense, artificial intelligence does not eliminate the need for universities. It clarifies it.

The institution was never meant to produce information. It was meant to produce people capable of dealing with information.

That task becomes more—not less—important as systems become more capable.

Europe, in particular, faces a choice. It does not lead in hyperscale AI infrastructure. It does not dominate global platforms. But it does have something else.

A tradition of:

• critical inquiry
• public responsibility
• institutional accountability

Europe may not compete on speed. But it can define the conditions under which speed is allowed.

Redrawing the boundaries

That requires a shift. From adoption to reflection. From capability to purpose. Not every technological possibility needs to become educational practice. Not every efficiency needs to be embraced.

Some forms of friction are worth preserving. Because they are not obstacles. They are the conditions under which thinking becomes possible.

Artificial intelligence does not ask universities to become more digital. It asks them to become more precise.

About what they teach.
About what they measure.
And about what they refuse to outsource.

What comes next

The university after AI will not be defined by the tools it uses. But by the boundaries it draws.

In the following essays, we examine where those boundaries begin to take shape—from the collapse of the take-home assignment to the changing nature of professional judgment in law, engineering and the human professions.

Artificial intelligence does not eliminate the need for universities. It exposes what they were always meant to do—and what they can no longer avoid defining.

This article is part of the series The University After AI, published in the Education & Culture section of Altair Media.

---

Photo by Anant Chandra / Unsplash