Are Universities Educating Students for a World That Is Disappearing?

Universities have traditionally prepared people for a world in which knowledge was scarce, expertise took years to accumulate and the ability to recall, analyse and apply that knowledge differentiated the graduate from everyone else. AI fundamentally challenges that model.

Knowledge today is becoming abundant. Analysis that once required days can increasingly be completed in minutes. AI can research, synthesise, write, code, model scenarios and generate plausible solutions at a speed no human can match. The educators and particularly universities response cannot therefore simply be to add AI modules to existing courses/degrees. The more fundamental question is, what should a university produce when intelligence itself is increasingly available on demand if not commoditised?

The answer may lie in moving education from the acquisition of knowledge towards the development of judgement, agency and human AI leadership.

Importantly, judgement itself is not uniquely human. AI is already remarkably capable of statistical, analytical and increasingly contextual judgement. Given objectives and sufficient information, machines will frequently make better decisions than people. The critical distinction here emerges higher up the judgement hierarchy. AI can determine what is likely to work and can optimise extraordinarily well once you define the objective but AI cannot determine which objective ought to take priority. That requires values, which come from people, institutions and societies.

Then there is Ethical judgement which is different, it involves balancing competing principles where there is no objectively correct answer. AI can explain arguments from multiple ethical traditions but it does not hold any of those values. It does not care which prevails, more worryingly it will reflect bias.

For me though accountability is perhaps the clearest dividing line, as I have written about earlier – The Accountability Gap at the Heart of AI | Beyond the Obvious. Imagine an autonomous vehicle choosing between two harmful outcomes. The AI can calculate probabilities and predict consequences but after the incident, society does not ask ‘What did the neural network believe?’, Instead it asks who designed/approved/deployed it and accepted the risk? This is where responsibility remains very human.

My view is us humans must remain responsible for deciding what is worth doing. That requires capabilities that conventional academic assessment often struggles to develop, notably balanced ethical reasoning, systems thinking, leadership, negotiation, empathy, creativity, resilience, accountability and the ability to make consequential decisions where evidence is incomplete and legitimate values conflict.

Educators should therefore consider becoming not merely curators of knowledge and universities centres of its transfer but laboratories of judgement. This suggests students might spend considerably more time confronting realistic ambiguity like corporate crises, cyber incidents, ethical dilemmas, geopolitical scenarios, boardroom negotiations or societal challenges. AI being realistically embedded within these exercises not prohibited, forcing students to determine when to trust it, when to challenge it and when human responsibility must override optimisation.This suggests universities should not simply teach judgement as though it were uniquely human. Instead, they should teach judgement above the AI layer.

Assessment would consequently change. Knowing the answer becomes less important when everyone has access to machines capable of producing one. Universities should increasingly evaluate how students frame problems, interrogate evidence, recognise assumptions, reconcile competing values, collaborate, communicate decisions,  defend their reasoning and maintain unbiased moral and ethical credibility. This represents a significant departure from today’s lecture, assignment + examination model biased by academic institutionalised proclivities.

The graduate of the AI era should not be someone trained to compete intellectually with machines. That is increasingly a futile and unwinnable contest. Universities have a more important opportunity to develop people capable of directing machine intelligence with wisdom, purpose and accountability.

Perhaps the future educational proposition can be expressed very simply … Teach students what machines can do, then develop in them what society cannot responsibly delegate to machines. The future premium in this lies less in making every decision personally and more in governing decisions made jointly by humans and AI.