What the Future of Higher Education Demands of Us

A dispatch from Cambridge

Jesus College, University of Cambridge

I’m writing this on the way back from a gathering hosted by the Noēsis Collaborative at Jesus College, Cambridge, where a group of higher education leaders, technologists, and philanthropists spent two days wrestling with a single question: How do we design higher education for human flourishing in the age of AI?

It was the kind of room where you’d expect sharp disagreements—people from different countries, different sectors, different politics. And there were real tensions. But one thing struck me: nearly everyone agreed that the liberal arts will be more important in the age of AI, not less.

Liberal education (the philosophy, not the ideology) is grounded in the idea that education should develop reasoning, judgment, and the capacity to think critically across domains. These capacities will undoubtedly become more valuable as AI takes on the technical execution of tasks that used to define what it meant to be skilled.

I agree. But I think the reason goes deeper than what most people are arguing right now.

• • •

The usual case for liberal arts in the AI era is essentially defensive: Don’t worry, your degree still matters. Employers will still value critical thinking and communication. That’s true enough. But it’s a thin argument—it reduces the liberal arts to a bundle of marketable soft skills and hopes that’s sufficient.

I think what’s actually at stake is more fundamental. AI is only as useful as your ability to evaluate what it gives you. And that ability—the capacity to interrogate an output, to know whether the premise is sound, to sense when something is missing—doesn’t come from a prompt engineering course. It comes from having built a base of knowledge and experience deep enough that you can actually judge.

I’ve written about this before: the realization I keep having with generative AI is that its output is only as useful as you can trust it. When I asked Claude to draw connections between George Orwell’s work and current events in higher education, the output was genuinely helpful to me—because I’d read the books. I had the context to know whether the analogies held up. Without that foundation, the response would have been impressive-sounding and uncheckable.

So the question for higher education isn’t just what skills will the AI economy reward? It’s what kind of knowledge do students need in order to use these tools effectively— with judgment rather than blind trust?

• • •

This is where experiential learning coupled with liberal education comes in—and where I think higher education has both an enormous opportunity and a deep structural resistance to seizing it.

I experienced this personally when I studied political science in college. It is one thing to read about political parties, political action committees, or how a bill becomes a law. It is quite another to work, as I did, as an intern on Capitol Hill and get a glimpse of how these concepts play out in real life. The experience was transformative—so much so that I wondered how it was even possible to actually master the concept without having spent time in the context. If you have ever had the experience of reading about something and then seeing it in real life, you know what I mean.

Now layer AI on top of that. A student who has used AI to summarize every major Supreme Court decision has information — but not the knowledge that comes from sitting across from a client whose life will be shaped by whether you bring their case forward. The judgment involved in that decision — weighing the strength of the legal argument against the cost to the client if you lose, considering what a ruling would mean not just for one person but for the precedent it sets — isn't something you develop by reading case summaries. It's something you develop by being close enough to the consequences to feel their weight.

We know this, actually. Play-based learning, scaffolded with formal instruction, is generally considered superior (the research on this is robust) to rote memorization for young children. As students grow into adolescents and young adults, the form of “play” changes, but the principle holds: learning is richer when it is connected to real life. Even the best teachers and the most vivid books cannot replace the experience of being in something yourself.

So if we know that the humanities will be increasingly important for developing students with the skills of curiosity, critical thinking, and judgment—and if we know that experiential learning is key to unlocking those skills—how might we develop institutions that marry these ideas?

• • •

This isn’t a new idea. Apprenticeship models, co-ops, and clinical rotations have done this for decades in fields like nursing, engineering, and social work. Researchers like Mary Alice McCarthy at New America have long discussed the benefits of “degree apprenticeships” that formally integrate work-based learning into higher education—drawing on models from the UK and pushing to make these pathways structural rather than ad hoc. More recently, Molly Kinder at Brookings has argued that as AI hollows out entry-level knowledge work, we need something like a medical residency model—structured, mentored programs where learning is the job itself. Some institutions are going further still: the London Interdisciplinary School, for example, was founded on the premise that real-world problem-solving—not disciplinary silos—should organize the entire educational experience. The question is why we treat experiential interdisciplinary learning as the exception rather than the rule.

Part of the answer is that it requires faculty to engage differently—not abandoning their research, but putting it in conversation with the people and contexts where it applies. I’d suggest that the best faculty already do this: they are embedded in research relevant to their field and in conversation with people who apply it. But systematizing that connection—making it structural rather than incidental—is a different kind of challenge. It asks institutions to let go of some of the boundaries they’ve built between the academy and the world outside it.

And it asks something of us as senior professionals, too. If AI is hollowing out what we used to think of as “entry-level” work, we need to reimagine what students are capable of when they arrive. That requires letting go of some of the paternalism embedded in how we think about readiness—the instinct that says you’re not ready yet when what it often really means is I’m not ready to let you try. It may mean being willing to let people in to the more challenging work earlier, giving them opportunities to test their knowledge in real settings.

• • •

I’ll be honest: I think what stops us from this kind of change is something a bit more uncomfortable than logistics. It’s the quiet conviction that we have nothing more to add than what we’re already giving. That if we can offload some of our work—to AI, to junior colleagues, to students themselves—there may be nothing else for us to do.

I don’t believe that. I’d be more inclined to feel threatened by that scenario if my to-do list wasn’t already too long, if the list of challenges I see in the world around me wasn’t already too weighty, if the time I felt I had to contend with them all wasn’t already too short. We do not have a shortage of hard problems. What we have is a shortage of capacity to throw at them—and a reluctance to build that capacity by letting others in.

I’ll write more about this in my next post. For now, I want to name it as the quiet force shaping both our action and our inaction.

• • •

I’m finishing this draft in the hours before the Artemis II crew launches for the moon—four astronauts strapped to the most powerful rocket ever flown with a crew aboard, heading farther from Earth than any human being has ever gone. Among them: the first woman and the first person of color to leave Earth orbit.

I remain convinced that the potential of the human race to innovate is boundless—hampered only by the limits of what we believe we’re capable of. And by our willingness to let more people in to the work of pushing those limits further.