Will 40% of U.S. colleges close? Michael Horn discusses the future of education
In this extended talk, Michael Horn—Harvard Graduate School of Education lecturer and co‑founder of the Clayton Christensen Institute—examines why technology has limited disruptive power in U.S. K‑12, how demographic and economic forces are driving a wave of college closures, and how AI, lifelong‑learning models, and work‑based pathways are reshaping post‑secondary education. He revisits the original disruptive‑innovation thesis, noting that K‑12 lacks true non‑consumption markets, while higher education faces oversupply, declining enrollment, and rising costs. Horn cites data that 15 % of U.S. colleges have already closed or merged, predicts 25‑40 % could disappear in the next five‑six years, and explains that elite institutions will survive while the “soft‑middle” tier is most vulnerable. He argues AI will first enable sustaining innovations—improved student‑success analytics and research productivity—and later power new, highly personalized learning institutions (AI‑driven bootcamps, apprenticeships, micro‑schools). He stresses that lifelong learning will become a societal norm, requiring short, stackable programs and experiential learning to keep pace with rapid skill depreciation. Horn also highlights the growing acceptance of alternative credentials (Google certificates, industry‑specific training) and bipartisan policy interest in expanding apprenticeships. Finally, he reflects on the pandemic’s role in accelerating parent‑driven school choice and micro‑school experimentation, suggesting that public districts could adopt similar personalized pathways. Throughout, Horn maintains an optimistic, yet measured tone, emphasizing that technology alone won’t transform education; purposeful business models and policy alignment are essential.
Chapters
There's no question there's great use cases of technology, but the opportunities to really be disruptive within the US K12 system are very few. Outside of the fringes on both the left and right, there's actually lots of agreement in the 80% around the education that we want our kids to have. 15% of colleges in the United States have closed or merged. Only a third of the the public believes in the promise of higher ed right now. I think we can be much more relevant, keep up with the pace of change in meaningful ways and not sort of fall back on the old pedagogical models that we've had. And I think we're going to be able to use AI to more rapidly spin up educational programs alongside those work based experiences to help make sure people don't have gaps in their learning. Lifelong learning won't just be a buzz phrase, it's going to be a way of how we work and exist in society. My name is Michael Horn. I'm a lecturer at the Harvard Graduate School of Education and the co founder of the Clayton Christensen Institute. There are no shortage of hot takes about the impact AI is going to have on business and society. Right now my own take is like anything else, it's not normative, it's how we use it, the business models we put it into and what we ask of it that's ultimately going to shape it. But I will say I'm an optimist and I think ultimately it's going to put much more power in the hands of individuals to be creative, to imagine things that they couldn't have otherwise done and create all sorts of new businesses, organizations, things that I think will be disruptive but will ultimately empower individuals and humanity over the long run. Clay Christianson, who was a longtime faculty member at the Harvard Business School, was known as the father of disruptive innovation theory. When we published disrupting class in 2008, we were really optimistic that technology properly implemented in new models of learning could really personalize learning to be sort of a tutor for every single child. And I think what we've seen in the years since is that there's no question there's great use cases of technology, but the opportunities to really be disruptive within the US K12 system are very few because there's not a lot of what Clay would have called non consumption areas where the alternative is truly nothing at all. In a way that really reinvents the system. I think we were right about a lot of the fundamental precepts, which is that we have an educational model that was never built to optimize the way students, individuals learn. It was built to sort them out at various intervals. It was built not for the knowledge economy, but for the industrial economy. And we really need to disrupt that model if we hope to leverage everyone's full human potential. I think we were right on all of that. I think where we were wrong was we maybe assumed some of the properties that AI frankly is bringing to the forest that didn't exist back then, that there would be systems that would learn from themselves over time and take in data from how individual students were doing and improve in ways that the education technology of that era just weren't doing. And I think we were right that we needed to fundamentally reinvent the model. Like technology is not good enough by itself to fundamentally create change. It's how it's used that really matters. But I think we were guilty perhaps of some magical thinking that somehow public schools that had been around for decades could just naturally incorporate that and make those really big shifts in their classrooms and sort of overthrow decades of regulation to make the changes to a truly student centered system in the time period that we hoped for. So in 2011, Clay Christensen was out in front of an audience and said, 50% of colleges will close or merge or go bankrupt over the next decade. As we dug into it a little bit more, really it's more like 25% of colleges and universities are running business models that just don't work and they're going to need to close or merge. And I think as we even dug into it more deeply, we said, really this is a two decade change. It's driven by demographic declines in the wake of the Great Recession. It's driven by a model that keeps adding cost but can't possibly charge people more. And it's driven by a model where they're starting to become substitutes in how we get our college educations and upskilling in the world of work. And what we're now seeing is that over the last 10 years alone, 15% of colleges in the United States have closed or merged over the past year. Almost one college every single week has closed. In public colleges and universities, we've seen whole systems consolidate in dramatic fashion in the states of Pennsylvania, Georgia, Connecticut, Wisconsin and more. And so we've really seen that we have overbuilt supply for dwindling demand over the past many years. And it's coming to roost right now. And the really interesting thing I think on top of that is that right now we're at peak demographics of the numbers of students graduating from American high schools, it's about to decline precipitously because starting in 2008, people had fewer babies. We have too many seats in colleges for 18 year olds with not nearly enough 18 year olds coming out. At the same time that a lot of people are skeptical about the value of, of an ever more expensive college or university experience and not sure, do I really get a return on the other side that makes it worth it? Under that weight, the 15% that have already closed, I think it's easy to imagine that we get to 25, 40% over the next five, six years. Here's the thing, right, which is if you're at a place like Harvard or Yale or Stanford or the state flagship schools or, you know, the Notre Dames of the world, you're fine. There's so much demand for limited seats at those universities. They've been able to charge high prices. They make a lot of money, frankly, from the research grants that come in. They're really research first use institutions. In many cases, they're incredibly important to the economies that surround them. Those sorts of institutions are going to be fine. It's really this soft middle tier where these institutions are very expensive. The expenditures underlying them continue to go up. They don't have strong brand names and they have trouble attracting students beyond 50, 75 miles of where they live. A stat that might surprise you is a decade ago, 40% of colleges and universities enrolled fewer than 1,000 students in this country. That's where when we see this hollowing out, there are 4,000 some odd colleges and universities in this country. That's where the hollowing out is really happening. It's not happening at the elite ends of our American university system. So as AI comes in, how's it going to change the existence of colleges and universities? I think if we're totally honest, we have no real clue yet. AI is still searching for product market fit in higher education, as Cal Newport, the noted computer scientist at Georgetown, said recently. But I think we can see a few things are starting to change. A lot of institutions are going to be able to use AI in what we would call sustaining innovations to improve the way they currently serve students. So student success measures to take in data and respond much quicker to their needs, to frankly create a more frictionless environment for students who are coming into college and university. The other piece where AI is making an immediate impact is in research. It's helping a lot of faculty members do research quicker, come up with different hypotheses, test them in Novel ways, crank out new papers and things of that nature, much faster, maybe much more readable in terms of teaching and learning. I think the jury is still out and my own hypothesis is that we're probably going to see a wave of new colleges and universities or new post secondary institutions, apprenticeship programs, boot camps, things like that, that are powered by AI, that create maybe more personalized teaching and learning experiences than we could have otherwise imagined. And it'll be really interesting to see what the impact on the market is of those innovations. So there's no question right now we're just at the beginning tip of AI changing the way we work. My sense is over the next couple years the changes will be modest, but over the next 10 years, 15 years, they're going to be pretty dramatic. Already it's said that people in technical fields, their skills depreciate basically every four years. So you need to be constantly upskilling and reskilling. I think AI is just going to accelerate a lot of those trends and the impact of that on colleges and universities. I think at least two seismic impacts. Number one, the need to be upskilling much more frequently throughout our lives is going to become paramount. And so having short burst, quick programs that we can just duck in, get more skills, duck back out of, into the workforce, much more seamless experience of work and learning, I think are going to be really important, but also huge opportunities for colleges and universities looking to serve more students. Because adults, all of us in the world of work, we're going to just need to be learning much more throughout our lives. Like lifelong learning won't just be a buzz phrase, it's going to be a way of how we work and exist in society. And colleges and universities have a big role to play there. I think the second thing that it's going to change is, is make experiential learning much more important to the learning experience. If AI is changing how we work really rapidly, keeping up with those changes, codifying them in say a textbook or an online course, it's going to be really hard. The moment you do it, it's going to be out of date, six months, a year later. But the more we can create really experiential opportunities for students to learn through cases, through simulations, on the job, things of that nature, I think we can be much more relevant, keep up with the pace of change in meaningful ways, and not sort of fall back on the old pedagogical models that we've had. The substitute products coming into the post secondary market, the low cost alternatives are really growing pretty rapidly. Google has its certificates, so you can learn IT skills. And not just certificates, but you can take them online through any number of partners that, you know, in the case of Coursera, serves literally millions and millions of learners. Today, it's not just Google, but you can name any IT company at this point, any tech company, and they tend to have a training program built around their ecosystem, Facebook, Unity, you name it. And so for technical skills, I think in many ways the job of colleges and universities would be how do we provide a wrapper around what they offer so that people don't just get the narrow technical skills, but they can get that certificate, that mastery and also something bigger with pedagogy and things that a college or university should be specializing in. Right, about how to actually teach and convey this information so people actually learn it. That's I think the first thing. The second piece of it is colleges and universities historically anyway, aren't just interested in the transmitting of technical skills, but also how you put that in a deeper context. What's often called the soft skills of learning, the critical thinking, communication, working with others through collaboration, things like that. And so I think colleges and universities, not just saying that, but really creating authentic places where people can collaborate and learn together through these different technical paradigms, I think is a place where they can really add a lot of value to the stuff that we don't know how to measure well, but we know is really important for human flourishing. In higher education outside of colleges and universities. I think we're at this moment where we're starting to see new forms of learning pop up. A lot of these are going to be work based experiences. So apprenticeships are becoming a much bigger point of emphasis and people and employers thinking, how can we leverage these to educate the future workforce without a degree, perhaps, but make sure we get really good individuals coming into these jobs to power our future success. A lot of people are looking at work based degrees. We have an alum of the school, Mallory Dwinnell, who is building work based degree program. So you get the best of the degree, but you get really practical experience on the ground and far more affordable because it doesn't cost a lot of money to educate people and they can be earning money in a job while they are learning as a core part of the experience. So I think we're going to see a lot of these alternatives to traditional college make much bigger headway in the years ahead. What we really have right now in the economy, I think is a signaling problem that traditional degrees have never been A perfect proxy for this person's ready for this particular job, but they're breaking down faster. And what's more, when we need technically skilled people in jobs, and frankly even more technically skilled because a lot of entry level roles are disappearing right now, people are going to need more technical skills, but they don't necessarily have degrees to get those. And so these alternatives to the traditional college experience that are much more ensconced in work itself, embedded with the skills and experiences people need to really succeed, I think is going to be the future. And AI is going to sort of accelerate that in two ways. One, because the changing nature of work is going to accelerate. Having people on the ground with you actually working on the problems, getting real experiences, understanding what the job entails and building the appropriate skills, I think that's going to be more important, not less in the future. And I think we're going to be able to use AI to more rapidly spin up educational programs alongside those work based experiences to help make sure people don't have gaps in their learning or have the right technical foundation to be able to be great at those jobs. I think the narrative around everyone needs a degree is already starting to shift. I think it's being driven by a few things. Number one, many, many, many families across the country are looking at colleges and universities and becoming much more discerning customers. They're saying roughly 38% don't complete a four year program within six years. They take on a lot of debt and there's no return. On the other side, if you get a lot of debt but no degree, a lot of them are looking and saying half of those who do graduate go into jobs that don't require the degree. And what the research is telling us is if you start in a job that doesn't require a degree, five years later, odds are very high that you're going to still be in a job that does not require a degree. At the same time, companies are starting to question, do I need to require degrees for all these roles? What's the real value of that? And so they're dropping degree requirements requirements, although they are still using degrees often to hire in practice. So we'll see how that shakes out in the years ahead. But here's the other piece of it that I think is really interesting. Gen Z coming up right now through the end of high school and graduating. More and more of them are going to vocational programs instead of college. More and more of them are saying, I'm going to take a gap year and figure out is College really? Right. For me, more and more of them are seeing the shortages, not just in the skilled trades, but also in tech fields and the like, and saying, huh, maybe there's a different path into those. And so you couple that with an environment where only a third of the public believes in the promise of higher ed right now. And it's not just among Republicans that the confidence is collapsing. I think Democrats are about 50% believe in college at the moment. And what I think you're seeing, though, is a shift in how both parties are actually thinking and talking about post secondary education. So anything after high school, which is to say, when President Obama was in office, this notion of everyone needs some college, everyone needs some sort of degree, was very ascendant in the Democratic Party in particular. And now I think you see a shift where both Vice President Harris and President, now President Elect Trump were talking about these college alternatives. And there's real bipartisan agreement that having a system of apprenticeships that's not just for, say, H Vac and plumbing, but also IT and salesforce jobs and Facebook marketing and things of that nature and healthcare and on and on, could be a real part of our education and workforce systems in this country. In many ways, while we see a big departure right now among the population, I actually think in an interesting way among policymakers, we see some convergence to saying, huh, we ought to look at these alternatives and think about how they promise pathways for all Americans, regardless of what really lights their fire. Obviously, Covid and the pandemic were brutal on a lot of families and K12 school systems, but I do think that there was an interesting silver lining in creating more room for innovation. Because while a lot of school systems, I think they were prisoner to their one size fits all mentality about how we educate and serve students and families from very different walks of life and very different learning needs at different times. What it did was empower parents to say, my neighborhood public school isn't the only way to get an education. And so I think what we've seen is that education is getting a lot less linear. And my own take is that disruption happens when there's either non consumption or people are overserved by the existing system. You might say, come on, Michael, how are people possibly overserved by our K12 schools? All we read about is how they're chronically underperforming or achieving. But I think for a lot of families, they're like, I don't really care about the football team or I don't care about this set of services, or I Don't need that. And so this micro school community that's extremely personalized, really rebuilding community in the way I think our public schools originally did and that are publicly funded in some states now, up to 12 universally. So I think for many families and students, that's creating a really interesting innovation pathway to drive disruptive innovation and drive personalization in some really interesting ways. And I think the question will be this, do we move to this fully alternative system or even better, do our public school districts start to say, huh, we could do that too. We can create alternative pathways in our own districts and create micro schools, schools within schools, different schooling hubs, if you will, where we really bring the best of community with personalization to empower people through many different pathways. And so in some ways, I'm actually more optimistic than I've ever been before because I think parents are really voting with their feet. Educators, interestingly enough, are most of the entrepreneurs behind these innovative forms of schooling. And together, I wonder, will they create a bigger shift that's more student centered in the long run? And more schools and more school systems will say, you know what, we can do that too. I think there's been a generation of kids, frankly continuing right now into the school systems that Covid just was devastating for. And I think a lot of students today are still extremely disengaged in school. We know enrollment in urban school districts has plummeted and continues to plummet. We know that chronic absenteeism remains incredibly high. And when you go into schools a lot of places, they just, they aren't very engaged. And I think this, rather than focus on the systems and the schools and the traditional structures, we need to be much more mindful of how each individual reconnects, re engages and something that's really motivating for them. And if that's sort of the entry level principle, then hopefully we can start to rebuild. But gosh, there's just millions of kids. This is an enduring fact of the pandemic that there's millions of people who didn't go to college, they didn't enter the workforce, and we don't really know what they're doing. How do we re engage those individuals, I think is a really tricky question for us right now. We're in a moment right now where baby boomers are leaving the workforce, they're starting to retire. And we know that employers consistently report they cannot find qualified people for large numbers of jobs. So we have this really weird moment in the economy that's more than a passing moment, which is that employers are keeping job openings open for long periods of time, really struggling to fill them. People are remaining stuck in work at the moment that they're not necessarily excited or passionate about, but they're nervous if they quit and how will they change jobs and find something different? And then there's large numbers of people who are disengaged, not in the workforce at all. And frankly they don't count in the unemployment rate because they're not actively looking for work. And so on one end you could say, well, okay, this is devastating to the individual communities, but I think it's devastating to the American economy writ large. And I just think if we can re engage those people who are on the sidelines with short burst programs that are low risk, that allow them to earn while they learn and get them re engaged with their hands, with their mind, something we could unlock a lot of growth in this country that we're completely discounting right now. What's always struck me about the education at a place like the Harvard Business School for MBA students is that yes, you delve into the technical details, how to build and read and analyze say a financial model, but you never stop there. You're always asking a question in a case of what would you do with this information if you were the CEO, what would you do as a manager confronted with this information? What's the question you really need to ask about this model? And so the way I think about it with AI is that our tools are going to change. But I think for students, having an underlying sense of the knowledge underpinning those answers that the models spit back at us is still going to be important so we can interrogate them, see where we agree or disagree, see where we think there's judgment calls involved. The data is not going to tell me what to do because it's backward looking. It's what's my theory of what this data is starting to show me and where's the world going that I need to be preparing for? And so being able to ask good questions and have great theories that allow you to act on the information, you know, ahead of time, I think is really going to be critical for leaders of businesses and society in the future.
