☕️ Good Morning! Thanks for reading Next.If someone forwarded this to you, get your own copy by signing up for free here.
In today's issue: Why the current approaches to AI on campuses don't work anymore; the process of closing down a college; how the college search is changing.
The June edition of “Next Office Hour,” my regular webinar series for higher ed leaders, will be focused on AI strategies for institutions. We’ll talk about how to go from running pilots on AI to a coherent, campus-wide approach.
I'll be joined in a conversation by:
Joseph A. Salem, Jr., Co-Chair of the AI Steering Committee, Duke University
Since generative AI arrived nearly four years ago, higher ed has followed a familiar arc: fascination, then fear, then a wave of policies and restrictions.
Most campuses have landed somewhere between a wait-and-see approach and a patchwork of siloed experiments. Neither is sufficient for the future, and the runway to move beyond pilots to scale solutions is getting short.
That’s the central argument in a new report, Designing the AI University: From Fragmentation to Foundation, published this week. After more than three dozen interviews with provosts, faculty innovators, CIOs, and employers, over the past year, a pattern became clear in my research: institutions know the urgency but haven’t yet made the shift from managing AI to designing for it.
Students are stuck in a confusing middle. The 2026 Lumina Foundation-Gallup State of Higher Education study captures the contradiction:
57% of students use AI daily or weekly in coursework
52% frequently encounter unclear or uneven course rules about it.
Nearly 3 in 10 say their institution isn’t adequately training them.
Usage also varies sharply by major. Technology and business students are the most frequent users, often because their departments actively encourage it. Humanities and natural-sciences students encounter far less consistency.
Entry-level work is disappearing. The job market isn’t waiting for campuses to catch up. There were 15% fewer entry-level and internship postings in 2025 than the year before, according to Handshake, while applications per posting jumped 26%. Employment for Gen-Z graduates in AI-affected fields has fallen 16% since late 2022, according to the Stanford Digital Economy Lab. Historically, new technologies favored junior workers. AI is running counter to that pattern.
As Erik Brynjolfsson told me, senior workers rely on “tricks of the trade,” tacit knowledge that helps them recognize when AI is wrong or incomplete. New graduates, by contrast, are competing directly with tools that can perform many of the baseline tasks that once defined their first jobs.
The crux: Entry-level roles used to function as training grounds as graduates learned by building slide decks, conducting research, analyzing data. AI is automating those tasks, and with them, the traditional on-ramp to the workforce.
From pilots to systems. The institutions making real progress treat AI not as a policy problem but as a design challenge that cuts across the entire academic enterprise.
The report outlines four strategies:
Set a floor. Every graduate should leave with baseline AI fluency: drafting and evaluating AI output, applying tools within their discipline, understanding the ethical stakes. Ohio State’s president Ravi Bellamkonda likened it to foundational knowledge in math and science. “We're going to ensure that there's a minimum that's met across the university,” he told me. “How much more than the minimum is up to the major.”
Build shared infrastructure. Treat AI like the library, not a departmental experiment. A common platform ensures every student, regardless of major or financial resources, builds on the same foundation.
Redesign the degree. Colleges need to help students “ladder over” the work AI is replacing. Co-ops, course-integrated industry projects, and work-based capstones need to move from the margins to the core of the curriculum.
Sequence learning. As outlined in the graphic below, AI fluency should build year over year: an introduction in year one, discipline-specific application in years two and three, integrated real-world projects by graduation.
Bottom line: The path forward doesn’t require a sweeping reorganization.
Campuses can start with a small AI design group to map where AI shows up across the student experience, identify gaps, and prototype new course designs.
Pair that with a lightweight governance structure so decisions can move faster than traditional shared governance allows.
Over time, that evolves into something more durable: a lean coordinating office that serves as connective tissue across the institution.
The goal isn’t centralized control. It’s enough alignment to move quickly as the technology changes.
✋ What to ask on the college tour about AI. If you’re touring campuses as a prospective student, asking about AI readiness is worth adding to your list. A few to consider:
Does the college have institution-wide expectations for AI fluency, or does it vary by professor and department?
How do faculty get support and training to incorporate AI meaningfully into their courses?
How does AI fluency build over four years, or does it show up only in isolated courses?
Is AI access provided to all students, or do some have to pay for tools on their own?
Can you give me an example of how AI is embedded into a required course, especially outside of the technology or the business school?
Has the college surveyed its own students about their AI experiences, and what did it find?
College closures generate headlines (just see the national media coverage of tiny Hampshire College’s closure last month).
But it’s what happens in the hours and days before the announcement that really matter to the fate of a college's property, its students and employees, and how alumni ultimately feel.
The problem is that college leaders often don’t think about that complicated mix of finance, operations, emotion, and logistics until it's too late, Doug Moore told us on the latest episode of the Future U. podcast.
🛑 Doug Moore is someone I've nicknamed “The College Closer.”I met Moore a few summers ago on a tarmac of a private airfield outside of Chicago, where he just landed his Cirrus SR20 single-prop plane after coming from a college he was about to close down.
The backstory. As Moore told me and Michael Horn, for years, he led the closure of industrial sites, mostly particleboard facilities, before a near-miss accident on a job site in Louisiana made him rethink his career. He started reading about the enrollment cliff, built a bench of specialists, and essentially invented a job that higher ed didn't know it needed. "What do you think it takes to close a college?" a friend asked him. "A lot more than it takes to close a particleboard plant," he told him.
Today, Moore calls himself something else: the grim reaper of higher ed. He's one of the few specialists in the country who helps institutions navigate the final stretch—not the decision to close—but everything that follows.
The cash problem nobody talks about. People assume colleges have money: endowments, donors, tuition revenue. But by the time Moore gets called in, the cupboard is usually bare. And closing a campus isn't cheap.
Just maintaining an empty building in a northern climate runs about $7 per square foot annually, roughly $500,000 a month for an average small campus.
Most college assets have been pledged as collateral against debt, leaving little to liquidate.
Moore's first call is always to the secured creditors. Blindsiding a lender you owe tens of millions of dollars to, he says, is the single biggest mistake colleges make.
🎽 The big drain on college budgets is athletics—and not big-time D1 sports as we often think, but all the small colleges that have started up new teams as a way to boost enrollment (see why in the clip below from Future U.).
There is no market for an empty campus. Urban campuses—like the University of the Arts in Philadelphia—can sell because the real estate has value independent of the college.
Rural ones are a different story. Alderson Broaddus in West Virginia, whose last appraisal was near $70 million, sold in bankruptcy for roughly $3 million including contents. Many rural campuses don't sell at all.
The emotional part. Moore is matter-of-fact about the logistics. He's less so about the human side.
Moore moved into the basement of Iowa Wesleyan for six months during its closure. He's thrown his own money into making sure graduations happen.
One of the first things he does after taking over is cease all collection efforts and track down students whose physical diplomas were being held over unpaid balances. "I've had kids break down in tears," he told us, "because mom didn't believe the kid had really graduated until we got them that document."
The bigger question: Moore predicts 400 colleges will close completely— abandoned campuses, not mergers—over the next seven to ten years. He thinks 20 to 30 more could announce by end of summer. The diagnosis is simple: "We have too many seats for not enough butts." The harder question is what happens to the communities left behind when the campus goes dark and no buyer ever comes.
🚨 For prospective families: One of the frequently asked questions I got on the road in my Dream School talks this year was how to determine whether a college is in trouble financially or might go out of business. I covered this subject in Chapter 9 of the book. Among the questions to consider:
Has net tuition revenue been growing, flat, or declining over the past several years?
Does the college have a bond rating from Moody's, S&P, or Fitch, and what does the outlook say?
Is the college drawing down its endowment beyond the standard 4% annually to cover operating expenses?
What percentage of students are athletes? If it's approaching 60-70%, that's a warning sign.
Are buildings well-maintained, or is deferred maintenance visible on a campus tour?
Has the college been subject to any heightened financial oversight from the federal government or its accreditor?
The number of colleges Harvard University selected as a peer when reporting their data to the U.S. Department of Education in order to get data to compare itself on various measures, such as enrollment and staff salaries.
🔎 AI Is Now a College Search Tool. Students are choosing colleges later, applying to more schools, and leaning on new tools to comparison shop, including AI. Wally Boston, a long-time higher ed observer breaks down the findings from a new survey of nearly 10,000 students from the high school class of 2025, which was conducted by EAB, a firm that helps colleges with their enrollment strategies. It found that the share using AI in their college search doubled in a single year, from 14% to 26%. That creates both an opportunity and a risk for institutions: colleges that aren't accurately represented in AI-generated answers may lose a student before they ever send a brochure. (Wally Boston)
📋 The FAFSA Comeback. Two years after the botched rollout of the simplified Free Application for Federal Student Aid (FAFSA) upended the college search for hundreds of thousands of students, the process of applying for financial aid is showing signs of recovery. The Class of 2026 hit an all-time high completion rate of 54.7%, with weeks still left before the traditional June 30 cutoff. Completions rose in every state. Still, even at a record high, barely half of seniors are completing the financial-aid form. (Higher Ed Dive)
🏛️ Remaking the Accreditors. An advisory committee last week signed off on the Trump administration's sweeping 173-page proposal to overhaul college accreditation, clearing the way for public comment. The proposal adds significant new mandates: accreditors would be required to police First Amendment compliance at public institutions, investigate research misconduct, and develop policies around "intellectual diversity." Critics argue the plan doesn't streamline accreditation so much as weaponize it, warning that making it easier for troubled colleges to switch accreditors could let bad actors evade oversight. (Chronicle of Higher Education)
If you like this newsletter and my books and want to support me, there are a few ways you can:
.png?upsize=true&t=1779742840263&width=500&composeType=play_button&overlayColor=%232f4254&overlayScale=0.3&name=Untitled%20design%20(11).png)
