The Collapsing Value Proposition of Higher Education—and How Universities Can Respond

Legitimacy, AI, and the Unbundling of the Modern University

This one hurt a bit to write, not gonna lie. I’ve been stewing on the (many) pieces of this re: higher ed for a while, and well, I guess it’s time to send it out into the world.

So, What Is the Single Greatest Challenge Facing Higher Education?

My answer to that question is straightforward in diagnosis but pretty complicated in implications. After all, I’m an (old) academic. Higher education has done many amazing things for many people, including me. As a first-generation college student who now teaches many students, including first-generation college students, the traditional model worked out for me almost exactly as advertised. I’ve truly been privileged to be part of so many students’ growth over the course of my career.

In many ways, it’s been an absolute hoot.

My short answer to the question up above, however, is that the single greatest challenge facing higher education is a collapsing value proposition. Universities must somehow (re-)demonstrate—clearly, empirically, and transparently—that they are worth the time and money, and do that in a world where the student pipeline is shrinking, the price is high, and the labor market signal of “degree = readiness” is increasingly noisy.

For most of the twentieth century, that value proposition was pretty obvious and universities were remarkably good at it. A four-year degree reliably signaled preparation for professional life and offered a pathway into the middle class. Institutions of higher education did not need to justify their model—they were the model.

Degrees also served as markers of social and political progress. Expanding access for women, racial and ethnic minorities, and first-generation students was not only an economic project but a civil rights project. Enrollment growth signaled inclusion, mobility, and institutional legitimacy. The university was not just a training system, it was a vital part of the postwar democratic promise.

That moment has passed, friends.

The question is no longer whether higher education matters in the abstract. The question is whether particular institutions, programs, and models justify their costs in a world where alternatives are proliferating and trust in institutions is weakening.

I’ll argue that universities are being “unbundled” along three dimensions—structural, temporal, and epistemic—and that AI, labor markets, and demography are accelerating that unbundling. The core problem amidst all that change, to my eye, is legitimacy.

The University Was a Bundle. The Bundle Is Unraveling.

In an excellent piece definitely worth a read, Steven Mintz recently argued that the American university historically offered a comprehensive bundle: academic instruction, residential life, credentialing, socialization, identity formation, and research. The model assumed full-time residential enrollment, a coherent four-year trajectory, and a curriculum combining general education with specialization.

This model emerged in a specific historical context: postwar prosperity, generous public funding, and a labor market that rewarded credentials. Institutional assumptions broadly matched student lives.

Today, they do not.

Most undergraduates are no longer 18–22-year-old residential students supported by parents and progressing linearly through a four-year curriculum. Many are older, working, commuting, transferring credits, or studying part-time. Universities have responded by layering flexibility onto inherited structures—online courses, transfer pathways, stackable credentials—while preserving the symbolic shell of the traditional model.

The result is a widening gap between institutional design and student reality. Universities continue to market a comprehensive residential experience, but increasingly deliver fragmented pathways stitched together across institutions, platforms, and time. The institution still sells a bundle that no longer describes the lived experience of most students.

Around 40–50 percent of U.S. bachelor’s degree recipients attend more than one institution (NCES Digest of Education Statistics, Table 326.30: https://nces.ed.gov/programs/digest/d23/tables/dt23_326.30.asp). Transfer and swirl patterns are now common features of the undergraduate experience (National Student Clearinghouse: https://nscresearchcenter.org/transfer-and-progress/).

So, the university bundle is unraveling, but pricing, branding, accreditation, and governance assume it is still intact.

Time Is the Hidden Institutional Technology

A second, less discussed pressure point is temporal. The modern university is not just a curriculum. It is also a system for organizing time: semester schedules, credit hours, cohort progression, and the tacit expectation that learning happens at a standardized pace.

Hollis Robbins has argued that AI-enabled personalization and tutoring could accelerate learning for some students, creating what she calls the “two-minute mile” problem: an educational system calibrated for a standardized pace confronted with students moving faster.

Even if only a minority of students accelerate, the implications are structural. Credit hours, accreditation standards, financial aid eligibility, and cohort-based progression all assume learning occurs at a roughly predictable rate. Time functions as a proxy for knowledge acquisition and maturation.

The four-year degree was never just a curricular bundle. It was also a temporal one. Universities sell knowledge and credentials, yes—but they also sell time. Time to grow up a bit, to explore interests, to build networks, and to signal readiness for professional life. Time is not just a side effect of college; it is one of the core products.

As learning becomes increasingly variable in pace (through dual enrollment, competency-based education, informal learning platforms, and AI-supported tutoring), time-based credentialing becomes harder to justify, harder to price, and even harder to interpret. A system built on standardized time struggles when learning no longer occurs on standardized timelines.

This is not speculative futurism. Students already arrive with substantial credits from Advanced Placement exams, International Baccalaureate, community colleges, and online platforms. The coherence of the four-year trajectory has already eroded.

AI simply accelerates a long-standing trend toward temporal fragmentation.

Universities as Epistemic Coordination Institutions

The modern university has never been merely a training center or credentialing platform. It has been a system for coordinating knowledge: producing shared standards of truth, expertise, and legitimacy.

For much of the twentieth century, universities solved a fundamental social problem: how a complex society decides what counts as knowledge and who counts as an expert. Accreditation, peer review, curricula, and degrees were not just bureaucratic artifacts; they were governance technologies for knowledge.

Universities did not eliminate disagreement, far from it, but they provided a shared procedural framework within which disagreement could be productively managed. That coordinating role is now under quite a bit of strain.

As standards fragment across disciplines, institutions, platforms, and publics, the authority of universities weakens. The result is not merely misinformation but incommensurability: groups operating with incompatible standards of evidence, expertise, and legitimacy. Universities once provided a stabilizing reference point. Increasingly, they are one voice among many.

So, yes, enrollment and funding matter, but the deeper meta-issue is epistemic governance, that is, who decides what counts as knowledge, and why anyone should trust those decisions.

The Unbundling of Epistemic Authority

Mintz’s analysis of the university as a comprehensive bundle clarifies this moment. The university bundled not just instruction and credentials, but authority over knowledge production, validation, and dissemination within a single institutional framework.

Knowledge production is now distributed across platforms, firms, think tanks, independent researchers, and AI systems. Validation is increasingly contested among these actors, and dissemination is instant and decentralized. Universities no longer control the full knowledge cycle in the way they once did.

This does not make universities irrelevant. But it makes their authority conditional and contested. Universities must now justify not just what they teach, but why they should be trusted as arbiters of knowledge.

Public trust data underscores this trend. Confidence in higher education has declined significantly over the past decade (Gallup Confidence in Institutions: https://news.gallup.com/poll/508169/confidence-institutions.aspx).

Whether one sees that decline as tragedy or overdue reckoning, it is a structural fact universities must now navigate.

AI as an Epistemic Stress Test

Public discourse often frames AI as either a solution to higher education’s problems or an existential threat. Both frames miss the deeper issue, really.

AI is not fixing universities’ knowledge problems. If anything, it is exposing how fragile those arrangements already were.

Universities have long claimed authority over knowledge, but replication crises, incentive distortions, and declining public trust have unsettled that authority. AI enters this weakened environment and institutionalizes those contradictions. It routinizes contested standards by embedding them into automated systems that inherit unresolved disputes about evidence, method, and authority. Further, it accelerates knowledge production without resolving those disputes, and it erodes universities’ historical control over what counts as legitimate knowledge.

In this sense, AI is best understood as a stress test. The challenge is not whether AI can teach calculus or write essays. The challenge is whether universities can credibly certify knowledge, skills, and judgment in a world where information is abundant and institutional trust is fragile.

AI will continue to drive the marginal cost of producing and synthesizing knowledge toward zero. The university’s historical business model—and its epistemic authority—were built on the opposite condition.

Further, AI is unlikely to become less capable or less accessible over time. Institutions built around information scarcity cannot assume a return to that equilibrium.

Human Judgment as a Positional Good

Robbins extends this logic by suggesting that as AI commodifies information and basic decision-making, human judgment becomes a premium “last mile” good. Algorithms can optimize broadly, but contextual, interpersonal, and normative judgment remains scarce and socially embedded—even if it is often biased.

Here is the crucial political economy mechanism: AI lowers the marginal cost of content delivery, but it does nothing to reduce the scarcity of social capital, gatekeeping, networks, and legitimacy. Those remain positional goods.

In higher education, this implies a familiar stratification. Mass institutions increasingly deliver commoditized content and credentials, while elite institutions ration high-touch mentorship, socialization, and gatekeeping. AI does not flatten this hierarchy, instead it clarifies which components of higher education are scalable and which remain scarce.

So this is less a technological story than a political economy story. Universities have always sold positional goods like networks, signaling, and socialization. AI threatens the commodity layer of the bundle while intensifying the scarcity of the human layer.

The likely result is a deepening bifurcation: luxury-good universities that sell mentorship and legitimacy, and mass institutions that sell credentials and content at scale. This has profound implications for inequality, social mobility, and elite reproduction.

The Labor Market Signal Is Getting Noisier

Historically, degrees functioned as labor market signals of cognitive ability, persistence, and socialization. That signal is now noisier…and it’s likely to become even noisier from here.

Employers increasingly question what degrees mean, eliminate degree requirements, or rely on alternative credentials. Burning Glass Institute and Harvard Business School document widespread “degree inflation” and the movement toward skills-based hiring (“Dismissed by Degrees”: https://www.hbs.edu/managing-the-future-of-work/Documents/dismissed-by-degrees.pdf).

Students increasingly accumulate credits across multiple institutions and platforms, making transcripts harder to interpret. AI further complicates assessment and authorship, challenging traditional approaches to evaluation.

So, college still pays off on average, sure. But variance across institutions, fields, and individuals is large and increasingly visible. Brookings and Georgetown CEW both show substantial heterogeneity in returns by major and institution (https://www.brookings.edu/articles/is-college-worth-it/ and https://cew.georgetown.edu/cew-reports/valueofcollegemajors/).

The burden of proof is shifting. Universities can no longer rely on generic claims about the value of college, they have to demonstrate outcomes.

The Demographic and Fiscal Squeeze

Demographics amplify these pressures further. The (likely too much discussed among academics in the last two years) “demographic cliff,” a decline in the number of college-age students following post-2008 birth rate drops, means fewer students competing for admission. WICHE projects a 12–15 percent decline in high school graduates by the late 2030s, with sharper declines in the Midwest and Northeast (https://www.wiche.edu/resources/knocking-at-the-college-door-11th-edition/).

At the same time, public disinvestment has shifted costs onto students. Student loan debt now exceeds $1.7 trillion and is approaching $1.8 trillion (https://educationdata.org/student-loan-debt-statistics). State funding per student has declined over the past two decades (SHEEO: https://sheeo.org/project/shef/). Public skepticism about college’s value has increased (Gallup-Lumina: https://www.gallup.com/analytics/644939/state-of-higher-education.aspx).

The university’s historical legitimacy rested on broad public trust and visible economic returns. Today, both are increasingly contested, and the institutions that once benefited from deference now face demands for proof.

Trust, Critique, and Living Inside the Knowledge Experiment

A deeper challenge is collective trust in how knowledge is produced and certified. For much of the modern period, universities coordinated standards for expertise and adjudicated disputes about truth.

Alternative knowledge producers—think tanks, platforms, independent researchers, corporate labs, and AI systems—now operate outside traditional university governance structures. The result is not simply misinformation or polarization, but incompatible standards for what counts as evidence and who counts as an expert.

This is a crisis of coordination. Who decides what counts as knowledge? Which methods are legitimate? Which experts deserve authority? And on what institutional basis?

Universities are no longer the uncontested answer to those questions. Nor is it obvious that they should be.

AI raises the stakes, obviously. By scaling knowledge production and blurring the boundaries between expertise and access, AI accelerates the decentering of the university as the primary arbiter of knowledge.

Universities are no longer managing epistemic change from above. They are living inside it.

Teaching, research, credentialing, and institutional authority are all being stress-tested at the same time—by technology, demographics, politics, and platforms that now mediate public knowledge. No one is steering this process, and there is no stable equilibrium to fall back on. Higher education is adapting in real time, under contested legitimacy, while the ground keeps moving.

The old settlement—where universities coordinated knowledge and society mostly deferred to the university’s authority—has broken down. What replaces it is still very much up for grabs.

What, Then, Are Universities Actually Selling?

For most of modern higher education’s history, universities sold a bundle: structured knowledge, a recognized credential, a social world, access to networks, and four years of institutionalized time.

Knowledge is now cheap and ubiquitous. Credentials are everywhere too, which means they mean less. Networks and socialization still matter enormously, but mostly for those already positioned to access them. And time, four (five, six…) years of semi-sheltered exploration, has quietly become one of the most expensive luxuries universities offer.

The remaining scarce goods are legitimacy and human judgment: mentorship, certification, socialization into elite norms, and access to networks that matter.

This is the uncomfortable core of the collapsing value proposition. Most students pay for a bundle whose most valuable components they do not fully receive at many institutions. Why? Because many institutions charge luxury-good prices while delivering commodity experiences.

So, how do we fix this?

Rebuilding the Value Proposition

The future of higher education will be shaped by demographics, labor markets, technology, accreditation regimes, and public trust. Institutions can redesign around outcomes and credibility, or layer accommodations onto inherited structures and hope legitimacy holds.

It seems to me that rebuilding the value proposition requires a few things, and this is just the beginning of the list perhaps.

Outcome transparency. Clear data on learning gains, employment outcomes, and mobility (e.g., U.S. College Scorecard: https://collegescorecard.ed.gov/).

Knowledge governance repair. Stronger norms of rigor, replication, methodological pluralism, and slower science.

Human differentiation. Honest articulation of what human mentorship and institutional judgment provide that scalable systems cannot.

Structural adaptation. Designing for fragmented pathways and variable learning tempos rather than pretending coherence still exists.

None of this is politically neutral. Outcome transparency threatens low-performing programs and cross-subsidies. Epistemic repair threatens entrenched incentive structures. Human differentiation threatens the fiction that mass higher education can offer elite-level mentorship at scale. Structural adaptation threatens accreditation, tenure norms, and tuition-dependent revenue models.

If these dynamics play out as expected, it seems to me that higher education is likely to bifurcate. Elite institutions will increasingly sell mentorship, legitimacy, and network access as luxury positional goods. Mass institutions will sell commoditized credentials and content at scale, with thinner margins and weaker signaling power. This is not just an educational story; it is a political economy story about stratification, mobility, and elite reproduction in an AI-saturated knowledge economy.

Educational stratification has always been a mechanism of political stratification.

The postwar democratic settlement assumed that expanding higher education expanded democratic inclusion. If universities become primarily mechanisms for elite reproduction and contested epistemic authority, that settlement collapses. At that point, the crisis of higher education is indistinguishable from a crisis of democratic legitimacy.

If you squint, you can see the rough shape of the deal universities need to offer: clearer outcomes, stronger epistemic credibility, and a more truthful account of what is genuinely scarce and worth paying for.

Conclusion: Legitimacy, Not Just Survival

The deepest challenge facing higher education is not technological disruption or demographic decline. It is legitimacy.

Universities must demonstrate that they are credible knowledge institutions, effective pathways to opportunity, and trustworthy stewards of human development. The old implicit social contract—pay us, trust us, and things will work out—no longer holds.

The value proposition now has to be rebuilt explicitly, empirically, and institution by institution. Some schools will do this well and thrive. Others will keep playing defense and slowly lose trust, enrollment, and capacity.

The era of unquestioned legitimacy is over.

What comes next is, in a very real sense, up to us.

---

If you enjoyed (hated) this piece, please share it with someone who you think would appreciate (hate) it. Thanks. :)

---

Also, if you’re interested in related writings of mine on these topics that explore some of these themes more deeply, here’s some links:

“Trust, Critique, and the Problem of Knowing Together”— https://kylesaunders.substack.com/p/trust-critique-and-the-problem-of

“AI Won’t Fix Our Epistemic Problems, It Just Reveals Them” — https://kylesaunders.substack.com/p/ai-wont-fix-our-epistemic-problems

“We Live Inside the Experiment Now” — https://kylesaunders.substack.com/p/we-live-inside-the-experiment-now

“Trust, Incommensurability, and the Limits of Intersubjectivity” — https://kylesaunders.substack.com/p/trust-incommensurability-and-the

“The United States Has Become a Low-Trust Society — and That Changes Everything” — https://kylesaunders.substack.com/p/the-united-states-has-become-a-low