What universities are for in the age of AI?

‘Beyond skills: reclaiming the purpose of higher education when intelligence is no longer uniquely human’, by Laura Maska, Aegean College

Generative AI is forcing higher education to confront a deeper question than assessment or academic integrity: what should universities cultivate in humans when knowledge-like outputs are no longer uniquely human?

Higher education is entering a phase in which its traditional advantage — the organised production, validation and certification of knowledge — is no longer enough. Generative artificial intelligence is not simply another educational technology. It is a redistribution of cognitive work. Universities are right to focus on assessment, academic integrity and AI literacy. But those are only first responses. The deeper question is harder, and more important: what must higher education cultivate in humans when intelligence itself is no longer exclusively human? [1–6]

For decades, higher education has operated on an implicit compact. Universities have equipped students with specialised knowledge, specialist expertise, and signals of competence that carry economic and social value. That implicit compact is now under increasing pressure. AI systems can already draft, code, summarise, translate, simulate dialogue and generate plausible interpretations across many domains. This does not make universities obsolete. It does mean, however, that the value of higher education can no longer rest primarily on the transmission of information or the production of first-pass outputs. The point of university education must shift from possession of answers to the disciplined exercise of judgement. [1,3,6]

From answers to judgement

This is why the central educational challenge is no longer knowledge acquisition alone, but cognitive integration: learning how to think with intelligent systems without surrendering thought to them. Students now work inside hybrid cognitive environments in which prompts, models, data, interpretation and human judgement interact. The key question is not whether AI helps them produce more, but whether it helps them understand more. That depends on capacities universities have too often treated as secondary: problem framing, epistemic awareness, critical evaluation, ethical reasoning, and the ability to decide when AI should not be used. [1–4]

Seen in this light, the real risk is not only plagiarism or overreliance. It is stratified learning. AI tools may be broadly available, but the ability to use them well is not evenly distributed. Some students are already learning to use AI as a cognitive partner — testing ideas, refining questions, checking assumptions and accelerating complex work. Others use it only at the margins, for convenience or surface support, or avoid it altogether because expectations remain unclear. The new divide is therefore not simply between access and non-access. It is between students who can convert AI into intellectual leverage and students who cannot. Left unmanaged, AI will not democratise higher education; it will deepen differences in confidence, guidance and disciplinary fit. [2,3,6]

Why generic AI literacy is too thin

This is why generic calls for “AI literacy” are too thin. AI is not used in the same way across disciplines, nor should it be. Engineering, law, design, psychology, business, medicine and the humanities differ in what counts as evidence, rigour, authorship and responsible judgement. A meaningful institutional response must therefore be discipline-sensitive. In some fields, students may need to learn how to integrate AI into modelling, prototyping or analytical workflows. In others, they may need stronger emphasis on interpretation, ethics, relational responsibility, voice, uncertainty and contextual judgement. The issue is not whether every student should use AI in the same way, but whether every programme is clear about what good judgement looks like in an AI-present environment. [1,2,4,5]

Equally important is the developmental dimension. Students do not arrive with uniform levels of readiness, and AI-related capabilities must be scaffolded over time. Early exposure may focus on basic understanding and ethical awareness, while later stages can emphasise critical evaluation, iterative use and integration into complex tasks. What institutions need, then, is not a generic AI bolt-on, but a staged approach to capability-building that links pedagogy, assessment and disciplinary purpose.

Assessment after first drafts

Assessment sits at the centre of this shift. Traditional assessment assumes that a finished product can stand as a reasonable proxy for individual learning. AI makes that assumption increasingly fragile. The answer is not a retreat into prohibition, nor a permanent arms race of detection. It is assessment redesign. Universities will need tasks that make thinking more visible: iterative drafting, oral defence, studio critique, reflective commentary, authentic application, collaborative inquiry and evidence of decision-making under conditions of uncertainty. The aim should not be to eliminate AI from academic work, but to ensure that its use remains intellectually accountable and educationally meaningful. Advance HE’s own conversation has been moving in precisely this direction: beyond detection, and beyond assessment understood narrowly. [4,5]

This has implications not only for pedagogy, but for institutional purpose. For too long, higher education policy has allowed “skills” and “employability” to become overly narrow organising terms. Those aims remain important. But an AI-rich world makes them insufficient. The most valuable capacities may be those least reducible to automation: judgement, interpretation, creativity, ethical discernment, civic responsibility and the ability to integrate different kinds of knowledge under real constraints. Universities should not respond to AI by shrinking their ambitions to technical fluency alone. They should respond by enlarging their human ambition.

A moment of choice

The real danger is not that AI will change higher education too much. It is that higher education will respond too timidly — layering AI on top of existing structures without rethinking what those structures are for. Institutions now face a genuine choice. They can treat AI as a compliance problem at the margins of the system. Or they can use it as a catalyst to clarify their educational purpose. The future

of higher education will not be decided by whether students use AI. It will be decided by whether universities can create the conditions under which that use becomes thoughtful, equitable and aligned with the deeper purposes of education.

The urgent question for higher education is no longer whether machines can generate knowledge-like outputs, but what universities must cultivate in humans when those outputs are no longer enough.

References

1. Yan, L., Greiff, S., Teuber, Z. et al. Promises and challenges of generative artificial intelligence for human learning. Nature Human Behaviour 8, 1839–1850 (2024). https://doi.org/10.1038/s41562-024-02004-5

2. Miao, F. & Holmes, W. Guidance for generative AI in education and research. UNESCO (2023). https://www.unesco.org/en/articles/guidance-generative-ai-education-and-research

3. Kasneci, E. et al. ChatGPT for good? On opportunities and challenges of large language models for education. Learning and Individual Differences 103, 102274 (2023). https://doi.org/10.1016/j.lindif.2023.102274

4. Hack, K. & Knight, C. Higher education in the era of AI. Advance HE (2023). https://www.advance-he.ac.uk/news-and-views/higher-education-era-ai

5. Advance HE. Generative AI: Beyond Assessment. Member Project 23–24. https://www.advance-he.ac.uk/membership/all-member-benefit-projects/Generative-AI-Beyond-Assessment

6. OECD. Artificial intelligence and education and skills. OECD topic page. https://www.oecd.org/en/topics/artificial-intelligence-and-education-and-skills.html

With over two decades of experience in Higher Education, Laura Maska has been with Aegean College since 2010. She has held multiple leadership roles, including Head of Institutional Development and Head of Admissions, and has served as Managing Director since 2018. Her academic background spans Business, Psychology, and Leadership and Management in Higher Education, equipping her with a broad perspective on institutional development and governance.