The year 2026 is an extraordinarily difficult time for any parent with a child thinking about college and career choices. Artificial intelligence is changing our world with terrifying speed, and, in the process … well, terrifying most Americans. Recent polls show that 70 percent of Americans think AI is advancing too fast; only 17 percent of adults expect its impact to be positive over the next 20 years. As for young adults in the workforce, 50 percent think AI’s costs outweigh its benefits—and unsurprisingly, several commencement speakers trying to get the class of 2026 excited about the technology were booed. Faculty members share the pessimism: 63 percent believe that those graduates are not well prepared for an AI-shaped workforce.
I understand these concerns. I’ve seen The Terminator, worry about many of the people developing the technology (I’m thinking of you, Elon, Sam, and Mark), and distrust a political class that seems clueless at best and craven at worst. But the alternative possibility is that, through wise governance, and after a rocky and disruptive period, we get AI right. The technology might finally give us the tools to crack the clean energy challenge and win the race against climate change, a race we are currently losing. It might unlock medical breakthroughs long considered impossible. It might usher in an age in which our basic material needs are readily and cheaply met. In such a world, the soul-deadening jobs of our era—from corporate law to back-office finance to countless bureaucratic roles—will be handled by AI agents. The body-breaking work of agriculture, construction, and mining will be done by robots. AI might invite us, at last, to rethink work itself: what we value in human labor, both socially and financially.
Just as we cannot know whether our collective future in the age of AI will be more dystopian or utopian (for the record, I’m betting on the latter), young people going off to colleges and universities right now are doing so at a time of enormous personal uncertainty. The new world they are entering is not yet defined. That makes it feel impossible to predict what kind of education will serve them best, or in what fields to invest their time and energy.
Anyone who claims to know with precision which jobs will survive the AI revolution—and in what form—is guessing. The honest answer is: We don’t know. What we can do is make intelligent bets. As someone who has spent two decades running universities and the past few years studying how AI is likely to transform them, here is my advice for playing the odds.
First, don’t bet on a specific job. Bet on a sector—three in particular.
The first is the care economy: health care, education, child care, geriatric care, social work, counseling, physical and occupational therapy, athletic and wellness training. These are jobs where AI will assist but where we insist on human presence because what makes them valuable is precisely the humanity of the person doing them.
The second is the creative economy: design, the arts, architecture, fashion, film, immersive game design, and the resurgence of handcraft. The University of Chicago economist Alex Imas, now DeepMind’s director of AGI economics, describes what is coming as “a post-commodity economy, where a growing share of expenditure goes toward goods and services whose value is inseparable from the human who provided them.”
As machines produce everything efficiently, what becomes scarce, and therefore precious, is the human touch. We won’t want robots comforting our children when they skin their knees on the playground, holding the hand of a dying loved one, performing onstage, driving an F1 car, helping us think through a complicated prognosis, or making ethical calls that impact us. AI might assist in all of those cases, but we will want humans in the lead, not merely in the loop.
The care economy and creative economy are not going away; if anything, they are where talent and status will migrate as AI colonizes knowledge work. Enrollment in health care programs is already booming. UNESCO believes that the world will need 44 million new teachers by 2030. A psychology major followed by a master’s in counseling is a sound strategy; psychology is one of those undergraduate degrees whose full value is unlocked only at the graduate level. Nursing, social work, occupational therapy, architecture, education—these are not consolation prizes in an AI economy. They might be the most consequential careers of the next generation (even if current Trump-era educational policies disfavor them).
The great news is that these jobs are meaningful and fulfilling when we pay people reasonably and don’t burn them out. They also make the world better, lifting people up and strengthening communities. Right now, many of these careers pay poorly. But in the future, when talent has been crowded out of the knowledge and technology sectors by AI and is migrating to human and relational jobs, there will be emphatic demands for good wages and better working conditions. The graduates of top universities who once flocked to consulting, finance, and technology-sector jobs will not tolerate the low wages today assigned to care work—jobs that, not incidentally, have long been done mostly by women.
There will also be a third category, smaller but critically important: AI management roles, in which humans work not “in the loop” but “over the loop,” maintaining judgment and control over AI systems woven into every corner of society. These will be among the most demanding jobs in history, requiring technical mastery, wisdom, moral seriousness, and the kind of judgment that cannot be automated precisely because it involves weighing things that resist quantification. This sector is likely to be smaller since a lot of AI “workers” will be overseen by AI “supervisors,” as is already happening. Also, in a department that today employs 100 humans, we will see the same productivity or more from an AI workforce overseen by one or two humans, three if each gets to go home after an hourly shift (the agents will work around the clock, of course). Given their demands, these jobs are likely to pay exceptionally well. But there will never be enough of them to offset the broader job losses AI will cause across most fields.
We should acknowledge that we will see new jobs invented, as we have with every previous technology revolution, and likely growth in some existing areas, such as cybersecurity. What is different and unknown about this technology is how many of those new jobs can and will be done by AI, and how many will require humans.
But choosing a sector is only part of the answer. The fundamental question is what kind of education prepares a young person for a world that is changing faster than any curriculum can track, a world in which the lower rungs of every organizational ladder are disappearing, in which what you know matters less than what you can do with what you know, and in which the distinctly human capacities of empathy, judgment, creativity, ethical reasoning, and the ability to make meaning are the ones that no algorithm can replicate.
That question leads us back to the university.
What to Ask a University
If I were accompanying a son or daughter on college visits today, my questions would fall into two categories: AI readiness, and the development of fully human beings.
On AI readiness, I would want to know whether the university has clear, coherent policies on AI use—not policies designed to prohibit or police, but ones that actively encourage students to develop fluency with tools they will be expected to master in their careers. I would want to see faculty who are themselves using the AI tools relevant to their fields, and assessment practices that have moved beyond proctored exams and blue books toward authentic, performance-based work that mirrors how knowledge is actually applied in the world. I would look for project-based learning, internships, cooperative education, and apprenticeships, experiences that ask students to do things, not merely demonstrate that they have memorized things or crammed for an exam. And I would want clear, honest answers about how the institution uses and protects student data, because the structures of trust in an AI world begin there.
These things are becoming table stakes. Every serious institution will have to deliver them. They are not what distinguishes the extraordinary from the adequate.
What distinguishes the extraordinary is something older and harder to measure: a commitment to developing fully realized human beings. This is not so much a new kind of education as a return to a very old and venerable one—the kind that preceded the industrial age and the utilitarian assumptions it baked into our universities.
I would look for a curriculum that teaches students how to think about their own thinking—one that builds metacognitive capacity, surfaces blind spots, demands intellectual rigor, and cultivates the habit of asking better questions. I would want to see genuine intellectual range, so students understand what it means to think like a scientist, like an artist, and like a historian—not because they will practice all three, but because each represents a different cognitive tool available to them when the world presents problems that don’t arrive pre-labeled by discipline.
I would look for intentionality around what we might call distinctly human capacities: empathy, ethical reasoning, leadership, and, at the right institution, attention to the spiritual dimensions of a life. I would want students to have serious opportunities to inhabit their embodied selves, through sport, movement, time in nature, practices of stillness.
Most importantly, I would want to see an institution that takes mentoring seriously, where faculty and staff are genuinely attentive to students, where the relationship between teacher and learner is understood as the beating heart of education, not an inconvenience to be minimized in the name of efficiency. The best colleges are not content-delivery systems. They are communities of formation.
Finally, I would want the institution to be clear about what it stands for. A prospective student might visit and decide that its mission and values are not his or hers—but the institution should know what they are and say them with confidence. Jesuit universities do this. The historically Black colleges and universities do this. The military academies do this. Women’s colleges do this. An institution without a clear sense of its own purpose cannot reliably produce graduates with a clear sense of theirs.
The Name We’re Afraid to Use
What I’ve just described is, of course, a liberal arts education. Reframed for an AI age, given new urgency by the coming transformation of work and society, but unmistakably rooted in the oldest and best tradition of what universities are for.
The irony is that when employers are asked what they want from new hires, they describe this education exactly, without calling it by name. Whether critical thinking, communication, problem-solving, or general social skills, the majority of employers want more than mere mastery of a discipline. Ninety percent say graduates benefit from interdisciplinary study and exposure to real-world problems. Eighty-one percent prioritize adaptability and flexibility. The Business-Higher Education Forum calls these “21st century workplace competencies.” The Association of American Colleges and Universities calls them “essential learning outcomes.” They are, in fact, what a liberal arts education has always aimed to produce.
The research is clear on one other thing: Parents want this education for their children. They want it passionately. But not if you call it “liberal arts,” a phrase that has accumulated so much political and cultural baggage that it has become almost unusable. Describe what it actually produces—critical thinkers, empathetic communicators, adaptable problem-solvers, ethical leaders—and parents respond with immediate recognition and relief: This is what I want for my child. This is exactly what I want.
Universities that understand this will quietly drop the label while doubling down on the substance. They are wise to do so. I’d also want to see them move past the old distribution or buffet approach to liberal arts—take one of these classes, two of those, perhaps one of these—and reimagine it for the demands of a new world. Brandeis University, for example, is doing exactly that: pairing its liberal arts core with required internships, skills credentialing, and dedicated career advising, backed by a $25 million investment from its board.
Today’s children will grow up in a world I cannot fully picture. The transition between here and there will be turbulent. There will be real suffering, real displacement, real political upheaval as the old economy is hollowed out and the new one takes shape. I worry about that. I worry about what it feels like to come of age in the middle of an earthquake.
But I also believe, with confidence, that the young people best equipped to navigate that transition will not be the ones who mastered the most in-demand technical skill of 2026. The graduates of Stanford’s famous computer science program, considered the surest bet for future success just two years ago, are finding this out the hard way.
Technological revolutions create what I call “flips”: reversals of long-established hierarchies of value. We are already seeing these emerge with AI. For example, in the newsroom, writers and editors were always most valuable while fact-checking was left to interns and entry-level staff. In an AI world in which content comes in a torrent and is often false, the value of content creators is falling and the value of those who establish accuracy and truth is rising. In software development, we prioritized assembling great engineering teams, and getting the business analyst or subject-matter expert was almost an afterthought. As AI transforms software development, engineering is far less valuable and nuanced understanding of workflows and business processes is more valuable. As knowledge economy and technology jobs get displaced by AI, the reversal might very well be the expansion and elevation of human jobs in the care or relational economy.
The graduates best prepared for that future will be the ones who were taught to think clearly and question honestly; who can meet uncertainty with curiosity rather than panic; who know how to listen to people whose lives are different from their own; who have a stable enough sense of who they are to adapt without losing themselves. They will have a robust cognitive tool kit, a big heart, and the resilience and self-awareness to navigate the one sure thing we know will be true: rapid change in an uncertain world.
That is what I want a university to give to students. Not a credential. Not a job. A self—capacious, resilient, humane, and prepared to flourish and be useful in a world we cannot yet see.
The rest will follow.
