In a 1963 lecture at Harvard, Clark Kerr coined the term multiversity to describe what the American research university had become in postwar America. At the time, Kerr was President of the University of California Board of Regents.
“In Kerr’s account, the modern university was no longer to be understood as a community of scholars united by a shared ideal of learning, but rather as a sprawling institutional conglomerate serving at once as a research engine, a job-training facility, a credentialing mechanism, a coming-of-age experience, and an incubator of the national technical elite,” wrote writer and historian Nils Gilman in his recent essay, “The University as We Know It Is Finished.”
According to Gilman, the multiversity model envisioned by Kerr was already becoming outdated in the 1960s because the postwar assumptions on which it was based had started to fray. Universities were becoming increasingly responsive to market demands. Instead of offering students primarily a broad education in the liberal arts, they increasingly viewed their mission as preparing students for the job market by teaching them the skills necessary to find good jobs after graduation.
At the same time, there was a significant shift from scholarships and grants to student loans, transferring more of the cost of higher education to individual students.
“The net result has been the ever-upward-spiraling costs of undergraduate education, without a corresponding increase in the value of educational training or credentialling, and a loss of political support for the mission of universities,” noted Gilman. “These financial and political travails have heightened the contradictions between the disparate missions and functions of the multiversity.”
The End of the Multiversity Model of Education?
“Into this increasingly unstable compound, add AI,” wrote Gilman.
“When a student can produce a plausible term paper in twenty minutes using Claude Opus or Google Gemini, what is the point of assigning term papers? When an AI tutor can explain any concept at any level of sophistication with infinite patience, what is the value of a lecturer reading from notes? When AI can ace most standardized professional examinations, what is a credential certifying? These are old problems that AI has made it impossible to ignore.”
Gilman argues that what comes next may be the disaggregation of the multiversity into its component parts.
“Actually, in many ways, it represents an opportunity to return to roots, in a classical model of education and in attentive pedagogical instruction. But higher education can only weather this period of disruption if it is clear-eyed about what is happening and moves confidently toward a new model.”
The multiversity articulated by Kerr was based on a familiar set of tools for educating and evaluating students: “the book, the lecture, the problem set, the written examination. In a pre-LLM world, these formats made cognitive demands of students that were difficult to simulate or shortcut.”
“That is no longer true. AI doesn’t just automate some of the tasks associated with these formats; it renders the formats themselves obsolete as instruments of either intellectual discipline or assessment.”
That may overstate the case. Books, lectures, problem sets, examinations, and term papers are unlikely to simply disappear. But AI raises serious questions about the roles they should play in education and, especially, whether traditional assignments remain reliable measures of what students themselves know and can do.
If AI can increasingly perform many of the intellectual tasks around which higher education has traditionally been organized, universities need to ask a more fundamental question: What should students learn to do especially well in a world where intelligent machines can perform so many cognitive tasks for them?
The Co-Curricular University
How should universities adapt to these major changes?
“The most popular present answer in certain administrative circles to this question is an emphasis on the ‘co-curricular,’ that is, on residential life and human connection as the university’s irreducible value in an age of AI tutors.”
Universities have long served two broad purposes. One is teaching and knowledge proliferation: educating students across a variety of subjects and preparing them for their working lives. The second is serving as a rite of passage into independent adulthood: living on one’s own, becoming part of a community that is learning together, and developing relationships with a wide variety of students and faculty members.
“Students’ own accounts of what matters most in college consistently center on relationships, belonging, and dialogue,” wrote Gilman. “The argument for residential education, for the ancient model of the Platonic Academy as gymnasium and garden as much as classroom, is stronger now than it has been in decades.”
“Any reimagining of the university in the age of AI must begin with an honest reckoning with what AI cannot do — and what therefore becomes relatively valuable precisely because AI can do everything else.”
This last point is particularly important. The challenge for universities is not simply to teach students how to use AI. They must increasingly help students develop the capabilities that complement AI — capabilities in which human judgment, experience, interpersonal relationships, creativity, and values remain essential.
Let me illustrate how humans and AI can complement each other with two concrete examples that I’ve written about in the past.
Decision Making: AI-Based Prediction and Human Judgment
Over the past several years, the need to align machine and human intelligence has been a common theme in AI research.
For example, in a series of articles on the economic value of AI, University of Toronto professors Ajay Agrawal, Joshua Gans, and Avi Goldfarb explained that decisions typically involve two major activities: prediction and judgment.
Predictions are generally based on information and analysis. AI is, in its essence, a prediction technology.
Judgment, on the other hand, involves determining the value of different possible outcomes and deciding what should actually be done. It often draws on subjective factors such as intuition, experience, values, and analogies with somewhat similar situations from the past.
“Judgment is the ability to make considered decisions — to understand the impact different actions will have on outcomes in light of predictions,” wrote the authors in an MIT Sloan Management Review article.
“Tasks where the desired outcome can be easily described and there is limited need for human judgment are generally easier to automate. For other tasks, describing a precise outcome can be more difficult, particularly when the desired outcome resides in the minds of humans and cannot be translated into something a machine can understand.”
As AI systems become increasingly capable of prediction, analysis, and information retrieval, judgment becomes relatively more valuable. Universities therefore need to help students learn not only how to arrive at answers but also how to ask good questions, evaluate competing alternatives, deal with uncertainty, and make thoughtful decisions when there is no single objectively correct answer.
Human-AI Augmentation in the Workplace
A second example involves how humans and AI can complement each other in the workplace.
In “The EPOCH of AI: Human-Machine Complementarities at Work,” MIT’s Isabella Loaiza and Roberto Rigobon argued that instead of asking whether machines are going to automate most jobs, we should shift our focus from the machines to the humans and ask different questions: How can humans and machines complement each other? And what human capabilities complement AI’s shortcomings?
Based on interviews with a wide range of experts, Loaiza and Rigobon identified five groups of capabilities that enable humans to perform work in areas where machines remain limited:
- Empathy and Emotional Intelligence: understanding other people’s emotions, needs, and perspectives.
- Presence, Networking, and Connectedness: developing relationships, trust, and human connections.
- Opinion, Judgment, and Ethics: exercising judgment and making decisions involving values and ethical considerations.
- Creativity and Imagination: generating new ideas and imagining possibilities that do not simply extrapolate from existing patterns.
- Hope, Vision, and Leadership: articulating goals, inspiring others, and helping groups move toward a desired future.
Together, these capabilities make up the acronym EPOCH.
This framework has significant implications for higher education. If AI becomes increasingly capable of performing routine cognitive work, universities should place greater emphasis on developing precisely those capabilities in which humans and machines complement one another.
The objective should not be to prepare students to compete against increasingly capable AI systems. It should be to prepare them to work effectively with those systems while developing the distinctly human capabilities that become more valuable as AI advances.
Reimagining the University in the Age of AI
How, then, is the university likely to change in the emerging age of AI?
“Start with the most obvious casualty: the term paper, as an assessment instrument, is dead,” noted Gilman.
“Written homework assignments were meant to push (and test) a student’s ability to produce a well-structured, coherently argued text. But this is exactly what LLMs do effortlessly and without demanding of the user any of the underlying cognitive work for which the traditional term paper was supposed to be a proxy.”
Among those capabilities is what Gilman calls sustained argumentative reason: “the ability to construct and maintain a complex argument across an extended piece of discourse, distinguishing claims from evidence, handling counterarguments, and reaching a defensible conclusion.”
The challenge is therefore not necessarily to abandon writing. Writing remains one of the most powerful ways to develop and clarify our thinking. That’s the reason why I’ve been writing my weekly blogs since 2005.
Rather, universities must find new ways to determine whether students are actually doing the underlying intellectual work — perhaps through oral examinations, classroom discussions, iterative writing, presentations, and assignments in which students must explain and defend their reasoning.
AI may have equally profound implications for the role of professors.
“If the post-AI university’s pedagogic value proposition is the formation of cognitive capacity in conditions that cannot be replicated on a screen, then the function and responsibilities of faculty members must also be reconceived,” wrote Gilman.
“It clearly no longer makes sense for professors to stand in front of a hall full (or, too often, only half full) of students delivering lectures. As a mechanism of information conveyance, AI can now provide the same at near-zero cost, tailor-made to the specific knowledge gaps of individual students.”
Instead, Gilman argues, professors should increasingly become interlocutors: helping students think through difficult problems, challenge assumptions, calibrate uncertainty, revise their views, and learn how to formulate goals rather than simply generate answers to questions posed by their professors.
This would represent a profound change in higher education. The professor’s comparative advantage would increasingly shift from conveying information to helping students learn how to think.
But such a transformation raises another difficult problem: scale.
“The central challenge for universities will be how to move toward this model at scale,” added Gilman. “The tutorial and seminar model is labor-intensive by design: a professor working as interlocutor rather than lecturer can engage only a fraction of the students she could previously reach from a podium.”
The required skills of faculty members would also have to change. Under the traditional research university model, faculty status and advancement have generally been based primarily on research achievement, while teaching has often been treated as a secondary responsibility.
But a university organized around dialogue, judgment, mentorship, and intellectual development would require professors with the pedagogical ability to understand individual students’ confusions and capabilities and help them reason through difficult questions. Research excellence does not necessarily produce those skills.
Reconceiving the professoriate would therefore require universities to reconsider tenure criteria, promotion incentives, faculty responsibilities, and perhaps even the relationship between research and teaching. Such changes would inevitably encounter resistance because they touch the fundamental structures and professional identities around which modern research universities have been built.
“Just as Kerr recast the University of California to match the liberal-technocratic imperatives of the postwar period, so do visionary college leaders today have an opportunity to remake the university to match the requirements of an economy that will be redefined by AI,” wrote Gilman. “Achieving this will be a generational project.”
I agree. But perhaps AI also gives universities an opportunity to reconsider a more fundamental question: What is a university ultimately for?
For much of the past several decades, universities have increasingly emphasized transmitting knowledge, providing credentials, and preparing students for employment. AI is now becoming remarkably capable at transmitting knowledge, tutoring students, generating sophisticated written material, and performing many of the cognitive tasks associated with professional work.
That doesn’t make universities less important. It may make their distinctly human functions more important.
The university of the AI era may need to place less emphasis on delivering information and more on developing judgment; less on producing answers and more on learning how to ask important questions; less on individual assignments and more on dialogue, collaboration, and intellectual community; and less on preparing students to compete with machines than on helping them develop the human capabilities that enable them to work productively alongside increasingly intelligent machines.
AI may therefore do more than disrupt the multiversity. It may force universities to rediscover one of their oldest purposes: helping people learn not simply what to think, but how to think — and, increasingly, how to exercise the judgment, creativity, empathy, and wisdom that intelligent machines cannot provide on their own.