The University's Quiet Bargain

Part Two of a Three-Part Series on AI, Education, and the Human Cost of Always Being Online

There is a moment many academics will recognise. A piece of student work arrives that is technically proficient, well-structured, almost impossible to fault; and entirely hollow. The conventions have all been observed. What is absent is the writer. The thinking. The alive quality of a human mind working something out.

After nearly two decades in learning and development and higher education across three continents, I have read thousands of assignments. Even the weakest submissions used to tell me something about the person who wrote them. I could see where they misunderstood a concept, where they wrestled with an idea, where they took a risk that did not quite work. There was a human being on the page. Increasingly, I finish reading and realise I have learnt almost nothing about the student behind it.

This is the new landscape of higher education. AI can produce competent academic writing within seconds, and universities are trying to decide what that means. Academics are encouraged to redesign assessments, treat AI as an inevitable part of learning, and judge students on how well they use it rather than whether they use it.

The adaptation is real. The question is what exactly is being adapted away from.

What the Struggle Was Actually For

The university has always been, at its best, a place of productive discomfort. Constructing an argument from incomplete evidence, holding a position under examination, failing a draft and rebuilding it, these were never inefficiencies in the learning process. They were the process. What emerges from that struggle is not simply knowledge. It is capability.

I do not remember my qualifications because of the certificates on the wall. I remember the struggle behind them. Postgraduate study completed while working. Doctoral research that challenged not only my intellect but my confidence. Drafts returned covered in comments, arguments that collapsed halfway through, ideas that took weeks rather than minutes to understand.

Over the past two years something has shifted in my classrooms. My students are no less intelligent than those I taught before. In many ways they are brighter, more connected, with access to knowledge previous generations could only dream of. Yet many seem more anxious than ever, and our conversations begin not with curiosity but with efficiency.

These are not separate conversations. The student reaching for AI as a shortcut and the university reaching for AI as a solution are responding to the same current, one that prizes speed and measurable output above almost everything else.

When institutions instruct academics to accept AI-generated work, they are not simply updating a policy. They are making a quieter decision about what a degree is actually for. If the output matters more than the process, the qualification becomes a signal of completion rather than capacity. The workplace has never been interested in what a graduate once submitted. It is acutely interested in what they can do.

The Pressures That Make This Happen

It would be unfair to frame this as a failure of institutional courage, though courage is sometimes what the moment requires. Universities face structural pressures that consistently point towards reduced friction. Satisfaction metrics. Completion rates. Fee income. The perception that being AI-forward signals relevance. Within that environment, accepting AI use can look like pragmatism, like inclusion, like meeting students where they are.

It may prove, over time, to look like something else.

The academics caught in the middle deserve acknowledgement. Many are deeply uncomfortable with what they are being asked to accept, uncertain how to push back without appearing resistant to change. The conversation in staffrooms is more troubled than official communications suggest. People are trying to hold on to something they believe matters, in an environment that makes it harder to say so.

The Equity Question Nobody Is Asking

AI tools are not evenly distributed in quality or accessibility. Students with premium subscriptions, strong digital literacy, and the confidence to prompt well extract far more from these tools than those without. The shortcut is not equally short for everyone.

A policy framed as progressive may quietly deepen the very stratifications it claims to address. First-generation students, international students working in a second or third language, students from under-resourced backgrounds, these groups may find themselves navigating an AI-shaped academic environment with fewer of the tools needed to do so well.

Equity requires more than permission. It requires investment, scaffolding, and an honest conversation about who benefits and who does not.

What Employers Are Finding

Workplace challenges do not arrive with rubrics. They require synthesis across incomplete information, communication with people who do not share your assumptions, the ability to persuade, adapt, and recover from being wrong. These are precisely the capacities built through the struggle AI now makes it possible to bypass.

Employers are noticing. Practical assessments, portfolio reviews, interview formats designed to surface what a CV no longer reliably signals, these are becoming more common. Graduates are arriving having been told they are prepared, only to discover the preparation was partial at best.

A student who has consistently offloaded difficult cognitive work to a machine has not simply taken a shortcut to a grade. They have missed the formation the grade was supposed to represent. That loss sits with them, not with the institution that awarded it.

The Conversation That Matters

Hope lives in the people who are refusing to stop asking difficult questions. The academics redesigning assessments to make thinking visible. The institutions experimenting with oral examinations, reflective journals, and portfolio-based learning. The students who sense something is being lost, even when they cannot yet say what.

AI belongs in education. Used with intention, it can deepen inquiry and extend access in ways that genuinely matter. The question is whether we are willing to be precise about what we want it to do, and equally precise about what we want to protect.

Education has always been, at its core, a relationship between a human being and a difficult idea. The technology changes. That relationship is worth keeping.

In Part Three of this series, we explore what a genuinely different relationship with AI could look like, for individuals, for institutions, and for the society we are still, together, in the process of building.