Colleges and universities are under acute pressure to prove their value in the form of graduate outcomes: Public confidence in higher education remains challenged and new federal accountability regulations require that program graduates earn more than they would have otherwise. At the same time, artificial intelligence has made the jobs market especially wily. Recent data suggests that AI has created a recession-like market for some new graduates. Yet the economy as a whole is still in the trough of AI’s J-shaped adoption curve, making its full impact harder to discern—all as the technology continues to evolve rapidly.
Cameron Sublett, senior director of innovation and incubation at FoundationCCC, which supports California’s community colleges, explained that job forecasting—with its direct implications for the postsecondary mission—has “always been a mad science,” and that AI disruption has added “so many more layers of complexity, confusion and noise.” Citing the economist David Autor, Sublett said that society is now in an “era of artificial intelligence uncertainty,” in which “history and secular trends no longer hold their predictive value.” Competing studies on AI’s labor-market effects regularly contradict each other, absent standard measures or methodologies, he continued.
Shawn VanDerziel, president and CEO of the National Association of Colleges and Employers, agreed that “the current employment environment is challenging for job seekers and employers.” That’s due to the uncertainty around AI, but also the broader economy and policy shifts, including on tariffs, he said.
Most Popular
All this compounds longer-standing financial challenges for higher education, added Dan Silverburg, founder and managing principal of the advisory practice Propel K20. No institution is immune to what’s been called higher ed’s polycrisis. And, explaining his own metrics for institutional viability using decades of data on college closures, Silverburg said the colleges most exposed in this new era are “exactly the tuition-dependent institutions whose value proposition rests on graduates getting jobs. For them, outcomes are not a compliance issue. They are the argument for existing.”
So how are colleges and universities trying to future-proof graduates when the future is so unclear and the stakes are so high? Experts say they’re doubling down on tested strategies—namely, experiential learning and durable skills development—while adding AI-native programs, reshuffling cabinets, requiring entrepreneurship and more. They’re also trying to help build the ecosystems that success requires. And in so doing, they might just be future-proofing themselves.
Historically, as a sector, “We wait for the market to shift, then go through a process that takes too long to adapt,” said Bryan DePoy, provost at Saint Leo University in Florida. “With skills and society evolving this rapidly, that posture is no longer viable. We need to be directly engaged with the workforce, crafting learning experiences in tandem with employers, not surveying them after graduates struggle.”
The tides are already changing. According to Inside Higher Ed’s 2026 Survey of College and University Chief Academic Officers, some 47 percent of all provosts agree that preparing students for a workforce shaped by AI has become a central organizing principle for academic planning at their institution. Still, colleges struggle with staff capacity around expanding work-integrated learning experiences—69 percent of provosts say this is a primary barrier, ahead of even employer interest or faculty buy-in—and just 20 percent agree that their institution has a coherent vision for how AI will change what to teach and how to teach it.
What Higher Ed Is Doing
Here’s what future-minded experts said higher ed can do more of to meet the moment.
Put career services in the president’s cabinet. Wake Forest University was an early adopter of this model, dating back to 2009. But Andy Chan, the university’s vice president for personal and career development, said it’s serving the institution well in the AI era. Chan—who once gave a TEDx talk on why “career services must die,” at least as a stand-alone concept—said that “if I’m not at that table, there’s no one really in the room sharing what’s happening in the marketplace around jobs and careers.” And when he raises an issue directly with deans and other vice presidents as peers, “those people will take me seriously and make it happen.” By contrast, Chan said, a more isolated career services director might ask for something and hear, “How important is this, really?”
Underinvestment in career services persists across higher education—and can show up in lower student return on investment down the line. According to recent NACE career services benchmarking data, the average career center employs 13 full-time employees, with an average student-to-staff ratio of 3,343 to one, and spends $102 annually per student.
Irma Becerra, president of Marymount University in Virginia, also changed the org chart at her institution: Career services leader Glenn Davidson is not only a member of the cabinet but also chief strategy officer. Becerra said that’s partly because internships—required for graduation at Marymount since its founding as a women’s college—demand relationships at the executive level. “He is the person that essentially represents me when being with companies,” she said of Davidson, adding that those conversations can help shape academic programs.
In a different kind of reorg, Saint Leo is executing a differentiated faculty model that recognizes and rewards industry engagement, DePoy said. “If we’re serious about preparing graduates for a changing economy, we have to be willing to change how we’re organized to deliver on that promise.”
Editor’s Picks
Scale experiential learning. Adam Weinberg, president of Denison University, described building “an all-of-campus approach to launching students quickly and successfully after their graduation.” The effort is anchored by the university’s Austin E. Knowlton Center for Career Exploration, where 30 career professionals work one-on-one with students—including to connect them with job- and internship-seeking alumni and partner employers. Denison funds internships to make sure all students can afford them and offers project-based micro-internships and consulting arrangements in which “students work to solve real problems for employers, present their findings directly and receive feedback,” he said.
Experiential learning, from internships to apprenticeships to employer-sponsored projects, is widely acknowledged as a can’t-go-wrong strategy. That’s due in part to employers’ ballooning work experience expectations for entry-level applicants. VanDerziel highlighted the value of paid internships, in particular, citing NACE’s own employer data showing that given two otherwise equal candidates, employers “will choose the one who has had internship experience.” Internship programs also deliver for employers “the best ROI in terms of recruiting strategies,” he said.
Scaling work-based learning, such as Northeastern University’s successful cooperative education, or co-op, model, is resource-intensive. Yet Chan cautioned that institutions can stall out imagining that “true experiential learning” requires an elaborate support infrastructure. Instead, he said, the real question is around its “minimum viable definition.” Chan said determining MVD starts by asking, “What are students doing now, already, that’s very close to experiential learning” that with more structure and self-reflection could be high-impact? Intermediary organizations can also help institutions connect career launchers with opportunities such as micro-internships, he said.
Zack Mabel, research professor and interim director of the Georgetown University Center on Education and the Workforce, said a missed opportunity in many places is simply thinking differently about campus jobs: “There’s a ton of work opportunities on campuses that are not intentionally created in such a way to facilitate the types of skill-building and relationship-building that we say is necessary for students to make a successful transition into the labor market.” He cited Arizona State University as one institution that’s found ways to restructure on-campus jobs for impact.
Silverburg pushed colleges to think about this as a curriculum-design problem: Audit every program for where work-based components exist, set expectations across the institution and count participation so that it’s visible to the provost and the board through annual reporting, he advised. He also cited Florida’s Valencia College and its large business advisory network as an example of an institution that treats employer engagement as a managed system.
At some institutions, professors—and their students—are also benefiting from AI-era faculty externships.
Build students’ social capital. As AI makes both applying to jobs and screening out applicants easier, students need to be meeting people who might know about opportunities, recommend them for roles and help them understand how an industry actually works, Mabel said. “In a world where more and more people are looking like viable competitive candidates, the only way you’re truly going to be seen is by being able to rely on some sort of network.”
Rachel Lipson, co-founder and scholar in residence at the Project on Workforce at Harvard University and author of the new book The New American Frontier: Job Training for the Next Technological Age, agreed that “the network power of a college or institution to help you connect is potentially more important than ever,” and that institutions focused on this are “well positioned to provide value to their graduates.” She pointed to engineering technology at Northern Virginia Community College as one program working to build students’ social capital via its strong career-readiness focus and industry ties. At Denison, the Denison Edge program offers career-credential programs taught by industry experts, and “Ask a Denisonian” connects students with specific career-launch questions that alumni can answer.
VanDerziel emphasized that students also need help learning to talk about and demonstrate their specific skills to employers, given the rise of skills-based hiring. And Mabel warned that as institutions lean into technology for simulated career-readiness experiences—another trend—these options are for building skills, not replacing the work that builds critical relationships and networks.
Focus on foundations and adaptability. Durable skills, including critical thinking, communication and creative problem-solving, are frequently mentioned in future-proofing conversations, but Chan argued against treating them as self-explanatory or requiring nothing new of institutions. “The old belief that ‘If you study this, you will get that’ was sometimes true, but not always,” he explained. “Now it feels like it’s rarely true.” And while this doesn’t mean the major is obsolete, he said, students need “a tool kit and a way of thinking that helps them navigate both [defined and emerging career paths] in the short term as well as in the long term.”
This tool kit also includes AI skills, but not as a bolt-on, said Weinberg: “We need to ensure students have the full range of durable liberal arts, AI and technical fluency, as well as the professional skills and habits required to succeed.” Saint Leo now treats “AI literacy as a baseline competency for everyone on campus, not just students,” DePoy added, underscoring institutional culture. “You cannot teach adaptability from behind.”
Silverburg warned that the durable-skills argument—while correct—only lands with employers when paired with “evidence that students have applied those skills to real work.” That’s why he believes that foundational skills, AI fluency and hands-on experience have to be built together. At the same time, he added, “Do not confuse AI for learning with AI readiness. One is about helping students learn better and the other is about helping them work like professionals, and they need different designs.”
Lipson, whose recent work focuses on emerging technical jobs, agreed with this “yes, and” take on durable skills: “There can be a lot of truth to the statement that people are going to need foundational skills that will transfer across their lifetime—and I don’t think that’s a strategy.”
The goal, said Susan Young, director of strategic initiatives at the Stanford Digital Economy Lab, should be “to help students succeed across a range of possible futures.”
Looking past graduation, some colleges are introducing career guarantees that provide temporary funding, access to graduate education or campus jobs to alumni who have not secured other opportunities. Institutions are also increasingly thinking about how to serve learners across their lifespan, including via the dual-speed model in which more traditional degrees exist alongside speedier upskilling units.
Add AI-specific programs and pathways. Antonio Delgado, vice president of innovation and technology partnerships at Miami Dade College, launched an applied AI pathway four years ago, just before ChatGPT became available to the masses. Colleagues asked at the time, “Are you crazy? You’re a community college. What are you doing playing with AI?” he recalled. “I was betting on a future where applied AI skills are needed.” And while early offerings struggled to enroll 30 students, the tiered pathway now enrolls more than 2,000—a third of whom are seeking an applied AI degree. The remainder are studying something else or working professionals seeking to upskill.
Delgado built the program to be stackable: AI awareness and AI practitioner certificates ladder up to a two-year applied AI associate degree and—through a 2+2-style arrangement—a possible bachelor’s degree. The college also this month launched what it calls the AI Elite Talent Hub, with a handpicked inaugural cohort of students who will be embedded with employer partners on real AI transformation projects, mentored jointly by faculty and company staff. The program ends with a paid internship.
“The moment you get to this level, you can work in a company,” Delgado said of talent hub students. “You don’t even have to wait to graduate—there are so many small and medium companies desperate for talent that cannot afford what the market is paying for AI.” All this creates a feedback loop that at once validates the applied AI pathway’s success and informs its future, he said.
Delgado is also co-leader of the National Applied AI Consortium, which has trained more than 3,000 faculty and staff members across the country and connected with companies including Microsoft, OpenAI, IBM, Intel and Google. And while the applied AI pathway is tailored to this moment, Delgado said he sees it as part of the workforce-aligned mission long held by two-year institutions: “It’s the role that community colleges play, to give you meaningful skills that five years from now might be completely different. But it gave you enough for you to get in” the door.
AI courses and programs at four-year institutions are growing rapidly, as well, from the University at Buffalo’s AI+ degrees to those affiliated with the University of Maryland’s Artificial Intelligence Interdisciplinary Institute at Maryland.
Students celebrate the launch of Miami Dade College’s AI Elite Talent Hub.
Miami Dade College
Treat entrepreneurship as a hedge—and a mindset. Saint Leo is working toward a requirement that every graduate will leave with “a presentation-ready business plan in hand,” not just a transcript, DePoy said. Planning is still underway, with the first phases launching next fall: Students will identify a problem worth solving in their first year, develop a concept and stakeholder analysis as sophomores, complete feasibility and financial analyses as juniors, and finish a full written plan and professional presentation as seniors. Because the requirement applies universitywide, DePoy continued, students will build plans “in the language of their own field”—so a nursing student’s will look different from a criminal justice major’s—along four possible pathways: venture, innovation within an existing organization, nonprofit initiative or professional practice model.
“Nobody can predict which job titles will exist in 2035, which means the institutions that survive and thrive won’t be the ones training for titles,” he said. “They’ll be the ones forming people who can ideate, adapt and create their own opportunity.”
Becerra said that Marymount is increasingly emphasizing both growth and entrepreneurial mindsets, to help students embrace failure as an opportunity, feel empowered to bring new ideas to the organizations they join and take ownership over their life outcomes: “You’re not just going to be successful on your first try, and how you pick yourself up and rethink your idea is very, very important.”
The goal should be to help students succeed across a range of possible futures.”
—Susan Young, director of strategic initiatives at the Stanford Digital Economy Lab
Watch (but don’t chase) labor-market data. Sublett, of FoundationCCC, called Sierra College a model of reducing uncertainty by studying the local economy, building relationships with regional employers—“hitting the pavement, shaking hands, inviting them to campus”—and using the U.S. Chamber of Commerce Foundation Talent Pipeline Management framework to build market-aligned programs such as surgical technology. This doesn’t translate to predicting the job market, he said, but it does mean operating in an informed manner locally, where job outlooks tend to be less hazy. Lipson added that noncredit programs can be potential incubators for market-aligned offerings, then transitioned to more permanent, credit-bearing status when outcomes prove strong.
Young, of the Stanford Digital Economy Lab, said it’s still too early to say with confidence which skills will matter more or less in the AI age, “and that’s one reason better measurement matters.” Her team’s AI Economic Indicators project—an extension of the headline-grabbing “Canaries in the Coal Mine?” paper on AI-era worker displacement—now tracks hiring trends in AI-exposed occupations, AI adoption rates and broader economic impacts, to complement other data with systematic measurement. And while higher ed leaders should be watching the data as it comes in, she said, they should do so cautiously: “If institutions underreact, they may miss important shifts in the skills and opportunities students will need to navigate an AI-enabled economy. If they overreact, they risk redesigning programs around trends that prove temporary, misunderstood or incomplete.”
With students, Young advised leaders “to be honest about the uncertainty.” AI is likely to change many occupations, but it’s unclear how, and how quickly, she explained. So rather than prepare graduates for a “single forecasted future,” she said, colleges and universities should help them build “strong foundations, learn new tools and technologies, and adapt as the labor market evolves.”
Build the ecosystem, because higher ed can’t do this alone. CEW’s Mabel was blunt: “It’s unfair to place the burden on colleges and universities singularly to address the challenges that this new wave of technology is introducing.” What’s required, he said, is “much closer partnership between our educational providers and our employers,” plus intermediaries who can help “bridge the long-standing cultural divide” and reconcile “at-times misaligned incentives” between the two worlds. He also suggested that colleges publicly quantify what they spend on maintaining relationships with employers to show that “getting to scale would require a level of investment that is just not possible to expect individual institutions to deliver.”
Silverburg argued that employers should “fund and co-design the experiences they say they want, not just show up to advisory meetings,” and that government can most usefully support the “intermediary layer between institutions and employers, which is where most partnerships break down.” He pointed to Germany’s dual-study programs and South Korea’s heavy investment in university-industry cooperation as models of scale and consistency. Nonprofits can help build shared employer pipelines that serve many institutions at a time, he added, while DePoy urged philanthropy to fund the advising, career coaching, employer engagement and other essential but “unglamorous machinery” that rarely attracts a named gift.
Lipson noted that U.S. co-ops declined in the 1990s, when federal funding for these outcomes-oriented programs dried up. She’s interested in how both state and federal policy can support gateways from education to work in this new era and highlighted Texas’s Skills Development Fund as a model where state training dollars only flow to employers who partner with an approved institution. Lipson argued for broader public investment in education-to-work bridges in a recent paper, saying they’re “no-regrets bets” to help early-career workers, “regardless of how AI scenarios unfold.”
Ultimately, DePoy said, “uncertainty is not a threat to higher education—it’s the case for it.”