After Scarcity: Rethinking the University in the Age of AI

AI does not eliminate scarcity in higher education. It moves it somewhere else.

Artificial intelligence is often discussed as though it creates a simple substitution.

• AI can write, therefore writing matters less.
• AI can explain, therefore knowledge matters less.
• AI can analyse, therefore expertise matters less.
• AI can tutor, therefore perhaps universities matter less.

I think this misunderstands what happens when a previously scarce resource becomes abundant.

Scarcity does not disappear. It moves.

And that may be one of the most important ideas for understanding what artificial intelligence means for universities.

This four-part After Scarcity series began with a deceptively simple question:

What becomes valuable when intelligence, explanation and content generation are no longer scarce?

Across four essays, I have approached that question through four lenses:

Power → Originality → Judgement → Purpose

Together, they suggest that AI does not diminish the importance of universities.

But it may force universities to become much clearer about what is genuinely valuable about them.

1. Power: when universities create knowledge but rent intelligence back

The first essay asked an institutional question:

Universities Built the AI Age. Why Are They Renting It Back?

Modern artificial intelligence did not emerge from nowhere.

Universities have spent decades producing research, researchers, intellectual property, datasets, graduates and ideas that underpin the contemporary AI economy.

That is precisely what universities should do.

• Knowledge should move.
• Research should generate social and economic value.
• But something peculiar can happen in the process.
• Knowledge flows outward.
• Capital and computing infrastructure accumulate around it.
• Commercial products are created.

And universities increasingly return as customers.

• The issue is not whether universities should work with companies such as Microsoft, Google, OpenAI or Anthropic. They clearly should.
• The more important issue is whether universities retain enough institutional capability and leverage to remain more than consumers of intelligence developed elsewhere.
• That makes questions of data ownership, open models, intellectual property, procurement, interoperability and internal AI expertise strategic rather than merely technical.
• The scarcity here is no longer simply access to intelligence.

It may increasingly be:

control over the infrastructure through which intelligence is accessed.

Read Part 1: Universities Built the AI Age. Why Are They Renting It Back?

2. Originality: when everyone has access to extraordinary intelligence

The second essay explored a different problem.

When Intelligence Becomes Abundant, Originality Becomes Scarce

Generative AI can make individuals more productive.

• Researchers can search, synthesise, draft, analyse and communicate more quickly.
• Students can generate ideas rapidly.
• Organisations can produce alternatives that once required hours of human work.

That is an extraordinary capability. But what happens collectively?

Recent research published in Proceedings of the National Academy of Sciences examined LLM involvement in US NIH and NSF research proposals and awards. Greater estimated LLM involvement was associated with proposals that were less semantically distinctive from recently funded research. At NIH, greater LLM involvement was also associated with proposal success and higher early publication output, although not with a corresponding increase in the most highly cited work.

That does not establish that AI causes unoriginal science, but it raises an intriguing systems question.

If millions of people use increasingly similar models, trained substantially upon the same accumulated knowledge, to determine what a good idea looks like, might we individually become more capable while collectively becoming more similar?

• The scarce resource may therefore shift.
• Producing a competent idea becomes easier.
• Producing an idea outside the established pattern becomes more valuable.

The question changes from:

Can you generate a good answer?

to:

Can you see something that everyone — including the model — has overlooked?

Read Part 2: When Intelligence Becomes Abundant, Originality Becomes Scarce

3. Judgement: when getting the answer is the easy part

The third essay moved from institutions and research to students.

AI Can Give You the Answer. But Do You Know Enough to Know When It Is Wrong?

For years we have treated access to knowledge as one of education's central challenges.

AI changes that. A student can now ask for an explanation of almost anything and receive a fluent answer within seconds.

But fluency creates its own problem. A convincing answer is not necessarily a correct answer, and the more capable AI becomes, the more subtle its mistakes can become.

Manu Kapur makes an important argument here: foundational knowledge becomes more, rather than less, important when AI can generate plausible expert-like answers so easily. Students need domain knowledge not merely to produce answers themselves, but to judge what AI produces.

That leads to what I think is one of the central educational paradoxes of generative AI:

The less expertise required to produce an expert-looking answer, the more expertise may be required to evaluate one.

This has major implications for assessment.

Perhaps universities should worry somewhat less about whether AI participated in producing an artefact and considerably more about whether the student can:

• interrogate it;
• identify assumptions;
• verify evidence;
• detect omissions;
• reject weak recommendations;
• compare alternatives; and
• ultimately take responsibility for a decision.

The scarcity moves again.

Not access to the answer.

Judgement about the answer.

Read Part 3: AI Can Give You the Answer. But Do You Know Enough to Know When It Is Wrong?

4. Purpose: what is the university for?

Those first three questions ultimately lead to the largest one.

The University in an Age of Abundant Intelligence

Universities developed in a world where sophisticated intellectual capability was scarce.

• Books were scarce.
• Experts were scarce.
• Libraries were scarce.
• Advanced analysis was scarce.
• One-to-one intellectual assistance was extraordinarily scarce.

AI changes many of those conditions.

Higher education is already adopting these technologies rapidly. EDUCAUSE's 2026 research found that 92 per cent of surveyed higher-education respondents said their institutions had some form of work-related AI strategy, while only 13 per cent reported that their institution was measuring the return on investment of AI tools.

That captures an important distinction.

Adoption is not transformation.

Giving everyone Copilot, Gemini or ChatGPT does not tell us what a university should become.

It merely makes the question more urgent.

If explanation becomes cheap, perhaps universities should concentrate more upon understanding.

If content becomes cheap, perhaps they should concentrate more upon creation.

If answers become cheap, perhaps they should concentrate more upon questions.

If artificial intelligence becomes ubiquitous, perhaps they should invest more deeply in the things that remain distinctively difficult:

• sustained attention;
• disciplinary expertise;
• intellectual independence;
• ethical reasoning;
• collaboration;
• real-world experience;
• curiosity;
• trust;
• judgement;
• wisdom; and
• relationships with other human beings.

The university's purpose begins to shift from the transmission of scarce information towards the formation of capable people.

Read Part 4: The University in an Age of Abundant Intelligence

The pattern across the four essays

Seen together, something interesting happens.

This is why I think scarcity provides a useful frame for thinking about AI and higher education.

The important question is not simply what AI can now do.

It is:

Once AI can do it cheaply, what becomes more valuable around it?

That is a very different question.

Universities should not compete with AI at the things AI makes cheap

There is a danger that universities respond to AI defensively.

• We detect it.
• Regulate it.
• Procure it.
• Train people to prompt it.
• Add an AI policy to the LMS.
• Introduce an AI literacy module.

All may be necessary, but none gets to the heart of the issue.

The more interesting opportunity is to redesign higher education around what remains scarce.

• A mediocre lecture that merely transmits information becomes harder to defend when a student can obtain a personalised explanation instantly.
• A brilliant seminar becomes more valuable.
• A generic essay becomes less useful as evidence of learning.
• A student defending a decision and explaining why they rejected several apparently plausible alternatives becomes more useful.
• A content-heavy course assembled around what students need to remember becomes harder to justify.
• An experience that builds deep disciplinary knowledge and then requires students to apply it to an ambiguous real problem becomes more valuable.

The excellent university may therefore become more valuable, not less.

But for different reasons.

Perhaps this is the real opportunity AI offers higher education

AI has forced universities into some uncomfortable conversations.

• About cheating.
• Assessment.
• Employment.
• Research.
• Intellectual property.
• Data.
• Costs.
• Power.
• Teaching.

Those conversations matter.

But behind them sits something much more consequential.

AI is forcing higher education to ask a question institutions should probably ask themselves periodically anyway:

What is genuinely valuable about what we do?

My answer after writing these four essays is increasingly clear.

• It is not simply access to answers.
• It is not content.
• It is not even intelligence alone.

The value of a university lies in creating environments where people develop the knowledge, judgement, curiosity, independence and relationships required to use intelligence well.

• AI doesn't make that mission obsolete.
• It makes the distinction clearer.

And perhaps that is the paradox at the centre of After Scarcity:

The more intelligence becomes abundant, the more valuable the deeply human capabilities surrounding intelligence may become.

• Power.
• Originality.
• Judgement.
• Purpose.

Those are not four separate AI problems.

Together, they may describe the emerging challenge for the university itself.

The complete After Scarcity series

1. Universities Built the AI Age. Why Are They Renting It Back? — Power
2. When Intelligence Becomes Abundant, Originality Becomes Scarce — Originality
3. AI Can Give You the Answer. But Do You Know Enough to Know When It Is Wrong? — Judgement
4. The University in an Age of Abundant Intelligence — Purpose

References

Kapur, M. (2026, July 28). AI makes foundational knowledge more important than ever. Times Higher Education.

Qian, Y., Wen, Z., Furnas, A. C., Bai, Y., Shao, E., & Wang, D. (2026). The rise of large language models and the direction and impact of US federal research funding. Proceedings of the National Academy of Sciences, 123(33).

Robert, J. (2026). The impact of AI on work in higher education. EDUCAUSE.