I don’t want to write a rant every time a technology mogul has a bright idea, but it’s tempting. The latest example is a $10,000 “AI college”, backed by names like Khan Academy, TED and ETS. It promises the same things that we’ve seen being touted again and again:

• personalised learning,
• a focus on competencies,
• and a model fit for the AI era.

It will be cheaper (well, for Americans), more flexible, more relevant. On paper, it ticks all the right boxes.

And yet, something about it feels less like a breakthrough and more like a familiar pattern in a new form. Because if you strip away the AI layer, this is another attempt to solve a complex educational problem with a model built around efficiency, scale and standardisation. George Ritzer had a word for that long before AI entered the picture: McDonaldization.

The idea is simple. Systems become organised around efficiency, predictability, calculability and control. Technology fits that logic almost perfectly. AI even more so. It allows you to scale, standardise, and automate feedback while still maintaining personalisation.

That combination is powerful in the eyes of the powers that be. It is also where things get tricky.

Because the real question is not whether this kind of model can work. It probably can, in certain contexts. The question is what happens when it becomes the default. Then it will be where an uncomfortable divide starts to emerge.

For those with fewer resources, education risks becoming largely technology-mediated. Structured, efficient, scalable. With quick feedback that is often generic. With less direct interaction with experienced teachers or mentors.

For those with more resources, the picture looks different. More human interaction. More coaching. Hopefully more nuanced feedback. But for sure more access to networks and opportunities that are not easily automated. In other words, we may be drifting towards a system where technology scales access, but human attention becomes the premium layer. Just like fast-food restaurants versus real food…

That is not a small shift. Because the real added value of education has never been about content delivery alone. It is in the interaction. In the feedback that is not only correct, but well-timed and well-judged. In the conversations that reshape understanding. And also in the social and cultural capital that comes with being part of a learning community. Those elements are hard to scale. And precisely because they are hard to scale, they become scarce. And because they are scarce, they become expensive.

So what is presented as democratisation can quietly become a very sad version of differentiation, even segregation. Of course, it would be too easy to dismiss this entirely. There are real problems with cost, access and relevance in higher education outside the European bubble I’m living in. Technology can help, especially where alternatives are limited. And some of these initiatives will generate useful insights.

But the economic logic behind them matters. If the main gain of AI is cost reduction, then the temptation will be to replace human interaction rather than to complement it. And unless there is a deliberate effort to do the latter, the system will not move towards more equality, but towards a new kind of divide.

Not between those with access to education and those without. But between those who get mostly technology, and those who still get people. And that might be the more important distinction.