There was a time when the case for university felt clear.
If you wanted access to expert knowledge, you went to the institution that held it. If you wanted a credible signal to employers, you earned the degree that proved you had done the work. If you wanted a better shot at opportunity, you followed the path laid out in front of you.
That path still exists. But...
What AI has done, particularly through large language models, is force a much more uncomfortable question into the open. What exactly is university for when knowledge is no longer scarce, when explanation is available on demand, and when a student can hold a meaningful conversation with a machine at any hour of the day? That question sits at the heart of this conversation, and Simon Brookes puts it plainly. Higher education is being pushed into asking hard questions about its purpose, its value, and the shape of the model going forward.
That pressure is real because the change is real.
The internet gave people access to information. Large language models reduce the friction involved in working with it. A student no longer has to rely solely on books, journals, office hours, or the occasional tutorial to get unstuck. They can ask for examples, ask the same question five times, interrogate a theory from different angles, and keep going until understanding starts to form. Simon describes this as having a kind of “super tutor” available all the time.
That changes the student experience.
It also changes the institutional challenge.
Because once a machine can write the essay, produce the references, format the citations and remove the usual giveaways, a written submission stops being reliable evidence of independent learning. Simon is very candid about that. He argues that undetectable AI-written coursework is already a live issue, that detectors do not solve it, and that universities are now being pushed back towards exams, oral assessments, presentations and professional conversations simply because traditional coursework has become so exposed.
This is where a lot of people jump to the wrong conclusion.
Some assume that university is therefore losing all value. Others assume the whole thing will settle down and little of substance will change.
Neither feels convincing.
The more serious interpretation is that higher education is being forced to rediscover where its value really lies. For a long time, part of the economic logic of university rested on access. Access to specialist knowledge. Access to research. Access to respected teachers. Access to a recognised credential. AI has not erased those things, but it has weakened the old scarcity model that supported them. Rob makes this point directly in the conversation when he describes how the historic value chain around knowledge and credentials is now changing.
And when scarcity moves, value moves with it.
The most important line in the whole discussion may be this one: the value is in the application of knowledge. That is the idea universities now need to build around. Knowledge remains essential, but knowledge on its own no longer carries the same premium when so much of it is instantly reachable. What matters more is the capacity to interpret, question, filter, combine, test and apply what is available. Simon returns to that point repeatedly, both in relation to students and in relation to the wider role universities play in research and society.
That shift has consequences far beyond the campus.
Employers have often used degrees as a proxy for something deeper than subject knowledge. In software, for example, we did not hire computer science graduates simply because they arrived knowing every framework or language in detail. We hired them because the degree suggested a way of thinking. It suggested first principles, abstraction, critical reasoning, problem solving and the ability to develop over time. That exact point comes up in the conversation. The degree acted as a signal for judgement and intellectual shape, not just immediate output.
That signal now becomes harder to read.
If AI can produce a polished artefact, then institutions need other ways for students to demonstrate how they think, how they reason, how they arrive at a conclusion, and how they respond when the machine’s answer is plausible but wrong. That is why the redesign of assessment matters so much. This is not only a question of cheating. Simon explicitly argues that universities need to look beyond academic integrity and think much more broadly about content, curriculum and what students actually need to learn to be ready for the world ahead.
That broader question is where things get interesting.
Because once you stop obsessing over whether AI can help someone complete a task, you start asking what kind of human capability becomes more valuable in a world full of machine capability. The answer that emerges in this discussion is judgment. Discernment. Critical thinking. Synthesis. Creativity. Communication. Ethical awareness. The ability to know when to trust an output and when to challenge it. The ability to decide where AI use is appropriate and where it erodes the thing itself.
That last point matters more than many people realise.
One of the most grounded parts of the conversation is the distinction between functional output and human expression. There are contexts where AI assistance feels perfectly reasonable. A report. A summary. A structured document serving a clear purpose. And there are contexts where people still want a real person to show up. Their voice. Their perspective. Their authorship. Their judgment. Simon and Kev both speak to this tension. They are trying to understand where AI use feels acceptable, where it feels lazy, and where it empties a piece of work of its human value.
That is not a side debate. It is part of the educational challenge.
If students are going to graduate into a world saturated with generated output, then one of the central things they will need is the ability to navigate that world intelligently. They need to know how to use these tools without surrendering all authorship to them. They need enough grounding to question outputs rather than absorb them blindly. Simon makes exactly this point when he talks about the importance of discernment and the danger of skipping the fundamentals.
This is also why the humanities question suddenly feels more urgent.
For years, humanities subjects have often had to defend themselves in economic terms against more obviously vocational or technical disciplines. But in a world where machines increasingly handle routine production, the traits cultivated through humanities education begin to look commercially significant in a different way. Interpretation, context, meaning, argument, critical reading, moral imagination, these are not decorative extras. Simon argues that the “humanities brain” may prove increasingly valuable as the workforce changes.
That does not mean technical education fades away.
It means technical education has to grow up.
The point is not simply to produce graduates who can execute a task the machine now performs faster. The point is to develop graduates who understand the system, can interrogate outputs, can set direction, can work from first principles, and can make decisions in ambiguous situations. In other words, people who can still think when the machine is doing a lot of the making.
University still has a meaningful role in that.
In fact, one could argue that its role becomes more important, not less. Simon makes a strong case that universities remain one of the best institutional environments we have for helping people develop those capabilities, while also supporting the research, networks, equipment, mentorship and long-term inquiry that serious knowledge creation still requires.
But that role cannot be defended with yesterday’s assumptions.
Students are paying a great deal of money. They are already questioning value. Simon is clear that institutions will have to adapt because students will vote with their feet if the offer no longer holds up. Some courses will change. Some may disappear. The requirement for pure knowledge acquisition will likely reduce, while the development of capabilities that machines cannot easily replicate will have to become more central.
That feels about right.
The degree is not disappearing tomorrow. The university campus is not about to become obsolete. The social, personal and developmental experience of higher education still matters, and Simon rightly emphasises that university has always been about more than content delivery. It is also a transformative phase of life. People learn how to communicate, collaborate, test themselves, develop independence and find out who they are becoming.
So will AI change university forever?
Yes, almost certainly.
The question is how deeply and how honestly institutions respond.
If universities treat AI as a temporary disruption to be contained, they will struggle. If they treat it as a prompt to rethink assessment, redesign curricula, develop discernment, strengthen application, and focus more deliberately on the human capabilities that remain scarce, they have every chance of becoming more valuable in the years ahead.
That may be the real challenge now.
Not whether university survives.
Whether it can explain, clearly and convincingly, why it still matters.