What can the higher education sector learn from the music industry about AI?

There is a wonderful moment in David Bowie’s now-famous 1999 interview with Jeremy Paxman when they start talking about the internet.

Paxman is sceptical, “It’s just a tool though, isn’t it?” Bowie is having none of it, “no, it’s not,” he replies. He describes the internet as an “alien life form” and predicts that the relationship between audiences and creators is about to change in ways we can barely imagine.

Looking back at that interview, what strikes me is not simply how accurately Bowie anticipated the internet’s impact. It is his recognition that technological change can transform the relationship between an industry and the people it serves. Mustafa Suleyman’s The Coming Wave explores a different but related challenge of how societies and institutions respond to powerful technologies whose consequences extend far beyond their immediate applications. For me, the connection is particularly meaningful because I have experienced the consequences of technological disruption first-hand.

I grew up in my family’s music retail business in Greece, established in the 1960s and eventually closed in 2005 as the economics of music were being turned upside down. I later completed my PhD on music piracy and researched how consumers chose between legal and illegal ways of acquiring music. Today I work in higher education, where I find myself watching another industry trying to make sense of a technology that challenges some of its most established assumptions and I am beginning to wonder whether we are making a familiar mistake.

The piracy problem

There is no revisionist argument to be made that music piracy did not matter.

It did. My own early research examined the decline in paid-for recorded music alongside illegal downloading and live concert attendance. Later, I studied consumers moving between multiple legal and illegal channels for acquiring music. What became increasingly difficult to sustain was the assumption that legal purchasing and illegal downloading represented two distinct groups of consumers. They did not. The same consumer could buy music, download it illegally and pay to see an artist live. Consumers were not simply choosing between legal and illegal music. They were choosing between different ways of accessing and experiencing it, and their behaviour was changing because technology had changed what was possible.

For years, much of the industry’s energy went into protecting copyright, pursuing illegal downloading and attempting to preserve revenues from recorded music. All understandable responses. But while the industry was fighting piracy, a much larger change was taking place. Music itself was becoming abundant.

Vinyl and CDs had bundled together several things. You bought an album even if you wanted only one song. You needed a physical object to access the music. Distribution was controlled. Ownership was central to consumption. Then suddenly, technology pulled these things apart. As access to recorded music became abundant, value began to emerge elsewhere.

A few years after that Paxman interview, Bowie became much more specific about what he thought this would mean for music. In 2002, he predicted that music would become “like running water or electricity” and that artists should prepare for much more touring because live performance would be one of the few truly unique experiences left. What interests me about that prediction is not that Bowie somehow foresaw the future. It is that he understood where value might move when music itself became abundant.

One of the propositions in my early research was that live performance could become increasingly important as recorded music revenues came under pressure. Look at the economics of music now. We pay a relatively modest subscription for access to a catalogue that would have been unimaginable when I was standing behind the counter in one of our record shops. Yet Taylor Swift’s Eras Tour became the highest-grossing tour in history. We made recorded music extraordinarily cheap to access while demonstrating just how much we are prepared to pay for presence.

But value did not move only towards experience. Something equally interesting happened to the physical object we had supposedly left behind. Vinyl returned – but with a different value proposition. It could never compete with streaming on convenience, choice or price, and it did not need to. What had once been the ordinary means of accessing music became tactile, collectible, intentional and scarce. We no longer needed the record in order to hear the music. That was precisely what allowed the record to become something else.

Streaming owns access. The concert offers presence. Vinyl offers something to hold, collect and value. The lesson is not that technology destroys value. It is that technology moves it.

What becomes more valuable

Higher education’s initial response to generative AI was understandably dominated by academic integrity and assessment.

The conversation has moved on. We are now redesigning assessment, developing AI literacy and exploring how the technology might transform teaching and learning. But I wonder whether even this more ambitious conversation risks missing the deeper disruption. The strategic question is not simply how universities should use AI. It is what becomes more valuable when everybody can.

The music industry asked how to stop illegal downloading. What it eventually had to confront was the much more difficult question of what people were still willing to pay for when access to music was no longer scarce. Our equivalent question is not simply what we should do with AI, but what the distinctive value of a university education is when access to knowledge is no longer scarce.

That question is much more uncomfortable. As Sal Khan’s work illustrates, if a student can ask an AI tutor to explain a concept five different ways at two o’clock in the morning, generate examples, interrogate an argument, translate difficult material, suggest ideas and provide immediate formative feedback, then some of what universities have traditionally provided becomes easier to access elsewhere.

I don’t think the answer is to compete with the machine. It is to work out what becomes more valuable because the machine exists.

Perhaps the scarce resource is the human being

Give two people exactly the same AI tool and they will not produce the same outcome.

One will accept the first plausible answer. Another will recognise that something is missing, connect the answer to an experience from another context, challenge an assumption, combine two apparently unrelated ideas and take the conversation somewhere unexpected. That difference is judgement. It is curiosity. It is creativity. It is experience. It is the ability to connect the dots.

This makes me think differently about the skills we should be developing in students. AI literacy matters, of course. But knowing how to operate the tool may rapidly become the least interesting part of being AI literate. The graduate advantage may lie in knowing what to ask, what not to believe, what deserves another question, and when an apparently sensible answer is solving the wrong problem. There is an institutional version of exactly the same skill.

A ticket for the HE concert

If AI makes some forms of knowledge and educational support abundant, what remains scarce?

An exceptional educator who changes how you think. A mentor who knows you well enough to challenge you. The argument after the seminar that sends you back to rethink everything. A research project where the answer is genuinely unknown. A professional network that opens a door. A community in which ideas collide. The confidence and judgement that develop because somebody expected more from you than you expected from yourself.

None of this means that face-to-face education automatically becomes premium. A mediocre lecture does not become valuable simply because it happens in a room. But exceptional human educational experiences might. And that raises some interesting questions about where the market goes next.

We may see a much clearer distinction between mass education – flexible, scalable, AI-rich and relatively inexpensive; premium education – deliberately human-intensive and experiential. Prestige education, where selectivity, reputation, networks and signalling remain part of what is being purchased.

The current move towards larger university groups makes this possibility more interesting still. Why assume that a university group of the future needs one proposition for one kind of student? A group could conceivably share technology and infrastructure while operating different educational brands across mass, premium and prestige markets.

There is, however, a problem

If human experience becomes the scarce part of education, who gets it?

This is where the analogy stops being an interesting exercise in market strategy and becomes a widening participation question.

There is a perfectly plausible future in which AI makes excellent educational content and support available to almost everybody, while intensive mentoring, small-group teaching, powerful networks, research opportunities and sustained contact with exceptional academics become increasingly concentrated among those able to access premium and prestigious institutions. We could be widening access to education while narrowing access to human education. That would be a strange definition of progress.

There is another possibility. Universities could use AI to make the abundant things genuinely abundant – explanations, routine support, practice, some administrative activity – and use the capacity released to invest deliberately in what remains scarce. More mentoring. More intellectual challenge. More opportunities to build networks and professional identity. More human attention where it changes trajectories.

For universities with a widening participation mission, that could become a much more compelling strategic proposition than trying to imitate institutions whose advantage rests partly on centuries of accumulated prestige. They could use technology deliberately, precisely so that human attention does not become a luxury good.

Back to Bowie

That is what I think we sometimes miss when we replay that Paxman interview. He did not possess a detailed map of the next 25 years.

His advantage was more fundamental. He recognised that technological disruption would change not only how music was distributed but where its value lay. He did not tell musicians to compete with digital music. He told them, in effect, to pay attention to what digital music could not replace. That distinction matters to higher education.

So yes, universities need rules. We need credible assessment and serious conversations about academic integrity. We also need to understand how AI can be integrated into teaching, learning and the curriculum. But these are defensive questions, and defensive questions alone rarely tell an institution where to go next. The more interesting question is where value moves when intelligence itself becomes increasingly abundant. If the history of music offers any clue, the answer will not be found by protecting everything that used to make sense. It will be found by recognising what has become scarce.

Perhaps the capability higher education institutions most need to develop now is the very capability we will increasingly need from our graduates. To use powerful tools without surrendering judgement to them. To bring creativity, experience and curiosity to what they produce. To connect the dots and recognise possibilities that were not obvious before. We will spend a great deal of time teaching students how to live and work with AI. But first, universities may need to demonstrate that they can do the same.

The music industry spent years trying to protect the business model built around recorded music. It took much longer to understand that the value of music had moved. Higher education should not need to learn the same lesson twice.