A recent essay in Comment magazine asks a question worth taking seriously: “What is the University For?” The author is a philosopher at Notre Dame who directs the university’s ethics initiative. The essay argues that generative AI poses two existential threats to higher education. Personalized AI dissolves the shared experience of learning by delivering a privately optimized encounter to each student. AI also replaces both the trusted gatekeeper and an open search with a single authoritative-sounding answer backed by no authority at all. The prescription is “counter-formation” through retreats, pilgrimages, common meals, and unplugged encounters organized around a values framework built on Dignity, Embodiment, Love, Transcendence, and Agency (DELTA).
The essay is eloquent and the concerns aren’t trivial. The argument rests on assumptions about universities, about AI, and about the relationship between the two that deserve closer examination than they typically receive.
The History That Isn’t There
Many arguments about the future of universities begin with a story about their past and the story almost always flatters the present institution. The typical version draws a continuous line from medieval Oxford to the modern research university as though it remains the same institution performing the same function across nine centuries.
However, the two institutions are fundamentally different. Medieval Oxford was a clerical institution operating under church authority. The curriculum was the trivium and quadrivium: grammar, rhetoric, logic, arithmetic, geometry, music, and astronomy. The pedagogical model was disputatio or formal oral argumentation governed by Aristotelian logic and Scholastic method. Students did not write expository essays. Students did not pull all-nighters in libraries. Students engaged in rigorous formalized disputation over set theological and philosophical questions, often under conditions of genuine material deprivation. The entire enterprise was ordered toward specific ecclesiastical ends.
The modern American research university has a fundamentally different genealogy. The roots of today’s university run through Humboldt’s German research model, the Morrill Land-Grant Acts, the post-WWII explosion of federal research funding, and the GI Bill’s transformation of higher education into a mass credentialing system. The university most of us recognize with tenure-track faculty pursuing original scholarship, undergraduate general education requirements, residential dormitories, and elective curricula is largely a twentieth-century invention. The library all-nighter with friends and takeout pizza is not a timeless feature of intellectual life. The all-nighter is a mid-to-late twentieth-century American experience belonging to a very specific material arrangement.
Telling the story of “the university” as a single continuous tradition stretching back to the twelfth century is not doing history. The narrative is institutional mythology. The romanticism does structural work in any argument that depends on it. The romanticism makes the university appear timeless and therefore essential when the institution being defended is recent, contingent, and still in transition.
The argument here is not for the superiority of the medieval model. The modern research university has produced genuine goods the medieval university could not have imagined and would not have permitted, the inclusion of women and non-Christians, the freedom to pursue empirical inquiry without ecclesiastical censorship, and the expansion of knowledge into domains the medieval curriculum could not have conceived. The question is not which model to restore. The question is exactly what capacity both models have failed to sustain.
The most serious attempt to answer the question “what is the university for?” remains John Henry Newman’s Idea of a University delivered as lectures in 1852. Newman argued that the university’s purpose was the cultivation of what he called the “philosophical habit of mind” or an integrative intellectual capacity that could perceive connections between fields of knowledge and exercise judgment across domains. Newman’s argument was already a response to the increased fragmentation and specialization of knowledge in the nineteenth century. Newman was not describing a stable institution. Instead, Newman was making a case against trends he saw undermining the university in his own time: the reduction of education to professional training, the loss of theology as the architectonic discipline that ordered all other knowledge, and the retreat of the university into specialization without integration. Newman’s university was an ideal he was fighting for and not a reality he was reporting.
The irony is that the contemporary university largely abandoned Newman’s project. The modern research university is organized around ever-narrower specialization with departments that barely communicate and disciplines whose practitioners cannot read one another’s journals. The integrative intellectual capacity Newman called for is precisely what many universities stopped cultivating.
Generative AI arguably embodies this capacity now. Foundation models are trained on the comprehensive breadth of human knowledge. Foundation models can model connections between fields that no individual specialist could perceive exercising a form of synthetic intelligence that moves across disciplines, languages, and traditions with a fluency that the siloed modern university cannot structurally match. The philosophical habit of mind that Newman fought for and the university failed to deliver may have found its most capable expression not in the institution he championed but in the technology it now treats as an existential threat.
The implications extend well beyond the university itself. Decades of development studies scholarship have documented what happens when siloed disciplinary knowledge is applied to complex human problems: economists who cannot see culture, political scientists who cannot see ecology, and technologists who cannot see community. The pluriversal critique of development articulated most fully by scholars like Arturo Escobar holds that the imposition of knowledge from a single disciplinary vantage point, no matter how sophisticated, reproduces the very extraction the imposition claims to remedy.
Genuine development requires holding multiple knowledge traditions and multiple ways of knowing in productive dialogue rather than subordinating all of them to a single framework. The university organized into non-communicating departments, each guarding its methodological territory, structurally reproduces the siloing that development scholars have identified as the root of failure. Foundation models by contrast can hold multiple frameworks in tension, move between disciplinary languages, and integrate perspectives that the departmental structure of the university today typically keeps separate. Some universities have recognized this problem and developed interdisciplinary programs that try to bridge departmental boundaries. The attempts are real but they remain exceptions within an institutional structure whose default logic is specialization and whose incentive systems still reward disciplinary depth over integrative breadth. The capacity for synthetic integration across knowledge traditions is not a trivial feature of the technology. The capacity is the architectural embodiment of the interdisciplinary and pluriversal inquiry that both Newman and development theory call for from entirely different starting points.
Even the canonical defender of the university here in the essay recognized that the institution required a unifying intellectual vision in order to function and that such a vision was already under threat. The question is whether the contemporary university has any such vision left to offer and whether the university can recognize that the integrative intelligence it claims to protect may already exist outside its walls.
The Revolution Inside the Gates
A further problem arises when the university is treated as the natural guardian of truth. The actual intellectual character of the contemporary humanities scene in today’s universities has been shaped decisively by a post-1960s turn toward critical theory, poststructuralism, and various Marxian and neo-Marxian methodologies. The post-1960s turn is not a footnote in the history of higher education. The turn represents a fundamental reorientation of what universities understand themselves to be doing.
The classical and medieval university sought to transmit and extend a received tradition of knowledge ordered toward transcendent ends. The post-1960s university across much of its humanities and social science apparatus explicitly positioned itself against received traditions and treated them as ideological formations requiring deconstruction. The pattern is most easily seen when remembering that the most cited philosopher across all the disciplines in today’s universities continues to be Foucault, from accounting to history and the social sciences. The hermeneutics of suspicion and the systematic interrogation of power and privilege embedded in claims to truth became not one method among many but something closer to the default intellectual posture of today’s university.
The result is a peculiar situation. An argument that the university must serve as the bulwark of truth against AI’s sophistry lands in an institution whose dominant intellectual frameworks have spent sixty years arguing that “truth” is always already entangled with power, that knowledge institutions reproduce social hierarchies, and that claims to objectivity mask particular interests. The critical instruments the university itself has sharpened would if applied consistently identify the romantic idealization, the institutional self-interest, and the unexamined power dynamics embedded in any defense of the university as truth’s natural home.
An institution that has spent decades destabilizing its own truth claims cannot easily maintain its purported role as the guardian of epistemic integrity against a new technological challenger. The institution being defended is not the one being described.
Alasdair MacIntyre, who held the Rev. John A. O’Brien Chair of Philosophy at Notre Dame for over a decade, made this argument with devastating precision. In After Virtue MacIntyre argued that modern moral discourse had lost its rational grounding: moral claims that sound authoritative function as nothing more than expressions of preference because the philosophical frameworks that once gave them substance have been abandoned. In Three Rival Versions of Moral Enquiry MacIntyre went further and argued that the modern university is fundamentally incoherent because it houses competing and mutually incompatible traditions of inquiry without any shared rational standard for adjudicating between them. The university maintains the appearance of intellectual community while lacking the substance. MacIntyre’s diagnosis delivered from within Notre Dame itself suggests that a values framework consisting of five appealing words is precisely the kind of irrational gesture his work identified as symptomatic. The DELTA framework sounds like a substantive moral claim but without a tradition of rational inquiry that would give each term determinate content, the framework functions as an expression of institutional preference dressed in the language of philosophical authority.
The Irony of the Method
The irony cuts deeper than institutional history. The standard move in recent arguments about AI and education is Foucauldian problematization: take something that presents itself as neutral or beneficial such as AI as an educational resource and reveal it as a regime of power that disciplines subjects in particular ways. The method maps how the technology produces docile learners who do not realize what they have forfeited. The method identifies the institutional apparatus that benefits. The method names the discourse of “personalization” and “efficiency” that naturalizes what is actually a political reorganization of knowledge.
Foucauldian problematization is a powerful critical method. The problem is what typically comes next. Having deconstructed one regime of power the argument often installs another and exempts the replacement from the same scrutiny. The institution that will rescue students from AI’s disciplinary apparatus is itself a disciplinary apparatus with its own surveillance mechanisms, its own production of subjects formed according to institutional ideals, and its own power-knowledge arrangements. Foucault would recognize the move immediately.
Herbert Marcuse, whose One-Dimensional Man remains one of the Frankfurt School’s most penetrating diagnoses, would recognize something further. Marcuse’s central argument was that advanced industrial society absorbs all opposition into itself. The system does not merely suppress critique. The system incorporates critique, domesticates it, and renders it functional within the very order it claims to resist. Marcuse called this process the triumph of technological rationality: a logic of efficiency, management, and optimization so pervasive that even the language of resistance becomes another expression of the system.
The university’s response to AI is a striking case in point. An institution that claims to resist technological rationality and that responds by developing a branded values framework, launching an institute, and organizing its “counter-formation” into programmatic deliverables with acronyms is not stepping outside the logic of the system. The institution is performing that logic with remarkable fidelity. The DELTA framework is itself a managed, optimized, and packageable product. The framework is institutional rationality applied to the problem of institutional rationality. The retreat is scheduled. The pilgrimage has learning outcomes. The formation is assessed. The resistance is funded through the same grant structures and institutional mechanisms that govern every other function of the modern research university. Marcuse would observe that the university has not escaped technological rationality. The university has merely branded its captivity as liberation.
The failure here is not one of intention. The failure is one of intellectual resources. The university cannot think outside technological rationality because its dominant frameworks and the critical theory it absorbed after the 1960s were always already a product of the same modernity they claimed to critique. Marcuse himself understood this. His own analysis left him with no exit and only the hope for a “Great Refusal” that his own theoretical framework could not ground or sustain.
The prescriptions that follow the critique are rarely postmodern at all. Values frameworks built on dignity, embodiment, love, transcendence, and agency are robustly modern and indeed almost premodern claims about human nature. The argument needs modernity’s confidence in human nature in order to prescribe formation. The argument needs postmodernity’s suspicion of power in order to problematize AI. These two intellectual traditions are in profound tension with each other and the contradiction typically goes unacknowledged.
The greatest critical thinkers themselves eventually recognized this impasse. Derrida spent his later career returning to Augustine and to confession, gift, and the structure of promise. Nussbaum built her capabilities approach on Aristotle because abstract proceduralism could not deliver an account of human flourishing. Foucault in his final lectures at the Collège de France turned to the Christian ascetic fathers and Greek parrhesia because his analytics of power left him with no account of how a subject could be constituted rather than merely subjected. Each thinker hit the same wall: criticism can dismantle but criticism cannot build. At a certain point the refusal to build becomes its own form of domination and leaves existing structures intact by default because the resources for proposing alternatives have been discarded.
The post-1960s university excels at problematization. But, in an important sense, the post-1960s university has also largely lost the capacity for construction. Values frameworks and acronyms appear where some kind of a metaphysics ought to be in place.
The Modernization That Cuts Both Ways
Arguments about AI and education also tend to operate within a discursive structure that will be familiar to anyone who has studied the history of development economics.
The structure works as follows. A teleology is established in which encyclopedias gave way to search, search is giving way to AI, and each era supersedes the last. The present state is framed as deficiency: students are being “habituated” to atomized learning and the educational commons is “dissolving.” The transition is presented as inevitable and as something that ‘will be complete before we even have the chance to contextualize it.’ Technological adoption or institutional resistance becomes the prescribed response. The trajectory is fixed. The question is merely how to respond.
As I have argued elsewhere the word “modernization” carries a specific intellectual history. Mid-twentieth-century modernization theory prescribed a single trajectory from traditional to modern with Western industrial capitalism as the destination. The results included dependency, extraction, and permanent client relationships rather than the promised development. The discursive structure survived the theory’s intellectual collapse and migrated into technology where it continues to foreclose the most important questions: who defined the trajectory, whether the trajectory is real, and whether the people affected might shape the terms under which new technologies enter their institutions and communities.
The irony is that the modernization structure operates in both directions simultaneously. Technology companies frame AI adoption as inevitable progress and non-adoption as backwardness. University defenders frame AI as inevitable degradation and institutional resistance as virtue. Both positions share the same underlying assumption: one trajectory exists and the only question is how to respond. Neither position asks whether AI must dissolve the educational commons or whether that dissolution is a design choice made by specific actors with specific incentives. Neither position asks whether the people affected by these decisions might participate in defining the terms rather than simply undergoing a transition that others have declared inevitable.
The question that gets foreclosed in both framings is the question of AI’s nature. The question is whether AI is ontologically a sophist designed to hold attention rather than track truth or whether AI’s capability is fundamentally truth-tracking with foundation models becoming more capable precisely by modeling reality more accurately. The answer to that question determines everything that follows. If AI is a sophist by nature then retreat is the only option. If AI’s capability tracks truth then partnership becomes possible and the interesting question is not how to resist but how to design collaboration that serves inquiry rather than engagement metrics.
As I have argued in previous work the central confusion in most discussions of AI and education is the failure to distinguish between two very different things: narrow optimization that trains systems to maximize engagement metrics and foundation model development that builds capability through breadth and accuracy. The first produces manipulation. The second produces intelligence. Attributing to AI’s nature what is actually a corporate design choice about optimization forecloses the most productive questions available.
The Assumption of Permanence
Perhaps the deepest unexamined assumption in arguments about “what the university is for” is that the university is a permanent feature of civilized society. The question takes for granted that the university will continue to exist and asks only what role it should play.
The assumption is the institutional equivalent of nationalist thinking. The university like the nation presents itself as ancient, continuous, and natural when it is historically contingent and has been invented and reinvented many times under different social conditions. Universities have disappeared before. The great centers of learning at Baghdad and Córdoba did not survive the political upheavals that destroyed their supporting civilizations. European universities were suppressed, reorganized, and emptied of intellectual content during the French Revolution and the totalitarian movements of the twentieth century.
No law of nature says the university persists. The institution exists because specific social, economic, and political conditions sustain it. The institution either transforms beyond recognition or ceases to exist when those conditions change sufficiently. More than eleven centuries elapsed between Caesar crossing the Rubicon and the founding of the first colleges at Oxford. Rome produced legal systems, engineering, literature, philosophy, and administrative structures that shaped the Western world for millennia. Yet, Rome produced no universities. The institution simply did not exist for most of human history. Arguments that depend on the university being the answer to the questions AI raises have no foundation if the university itself is historically contingent and might not be the right institutional form for a world that is emerging in its present state.
The structural incentive deserves direct acknowledgment. The people making the case for the university’s indispensability are the people whose positions, institutes, and funding depend on the university’s continued relevance. The same critical methods that expose the self-serving character of technology companies’ “responsible AI” framing would if applied consistently expose the same dynamic in institutional defenses of the university.
The Disciplines That Aren’t Permanent
The assumption of permanence extends beyond the institution itself to the forms the university treats as essential to intellectual life: the research paper, the expository essay, sole authorship, and the prohibition on plagiarism. These forms are defended as though they were timeless disciplinary standards and permanent features of serious intellectual work. The forms are nothing of the sort and remain the technological rationality of a very specific era. Marcuse’s diagnosis here applies with particular force.
The research paper as we know it is a product of the modern research university, itself barely a century old in its current form. The expository essay has a far richer history as a vehicle for political and revolutionary thought, employed powerfully by Latin American scholars among others. However, today the essay remains an assessment instrument in the contemporary university, the product of twentieth century developments like mass higher education and standardized grading. Sole authorship as the normative model of intellectual production is a product of Romantic-era conceptions of genius filtered through copyright law and academic credentialing. Plagiarism as a disciplinary offense with the weight it currently carries is substantially a product of the technological capacity to detect it. Print culture made textual comparison possible. Digital instruments like Turnitin made detection routine. The prohibition scaled with the detection technology and not with any timeless principle of intellectual honesty.
Previous eras of intellectual life operated on entirely different assumptions. Medieval scholars compiled, copied, glossed, and synthesized without attribution as a matter of course because the goal was the transmission and extension of knowledge and not the certification of individual originality. The disputatio that formed the core of medieval university pedagogy was an oral, collaborative, and inherently social form of intellectual work. Classical rhetoric treated imitation of prior masters as the foundation of skill and not a violation of it. Quintilian’s Institutio Oratoria, arguably the most influential educational text of the Roman world, built its entire pedagogical method on imitatio: the student learned to write and speak by closely imitating exemplary models and incorporating their structures, phrases, and patterns of argument into the student’s own practice. Isocrates, whose school in Athens rivaled Plato’s Academy, understood education as formation through immersion in a tradition of discourse and not as the production of certified original outputs. The notion that a student’s primary intellectual product should be a solo-authored written text produced without assistance and assessed for originality would have been unintelligible to most of the university’s history and to the classical tradition that preceded it.
The contemporary university defends “academic integrity” as the institutional standard of one or perhaps two generations that were solidified during a period when print technology and credentialing requirements converged to produce a particular set of norms. These sociotechnical norms are not necessarily wrong but they are not permanent either. Treating these norms as though they define the boundary between legitimate intellectual work and fraud is precisely the kind of naturalization that Marcuse identified as characteristic of technological rationality. The system’s own contingent arrangements become invisible as arrangements and instead present themselves as the natural order of things.
The question of AI partnership makes this analysis concrete. An institution that treats human-AI co-authorship as inherently suspect and that frames AI-assisted writing as a form of cheating is not defending a permanent standard. The institution is defending the specific arrangements of a recent era and mistaking them for eternal principles. The anxiety about AI-generated text is structurally identical to earlier anxieties about the printing press, the typewriter, the word processor, and the internet. Each new technology disrupted the prevailing norms of intellectual production and each time the institution defended the old norms as essential rather than examining whether new forms might serve the underlying purposes of education, truth-seeking, and intellectual growth equally well or better.
The question is not whether students should use AI. The question here is what forms of intellectual work actually develop the capacities that matter: judgment, discernment, the ability to distinguish truth from plausibility, and the ability to collaborate across different forms of intelligence. Those capacities predate the research paper and will outlast it. Defending the forms rather than the capacities is a category error and only an institution trapped inside its own technological rationality would make it.
What Formation Actually Requires
The deepest problem with arguments that frame the university as the answer to AI is not any particular historical or methodological error. The deepest problem is the absence of positive content.
An institution that says it will offer “counter-formation” must answer the critical question: formation toward what? Values frameworks offer words like dignity, embodiment, love, transcendence, and agency. These words are aspirations and not a curriculum. The question is what texts students read, what arguments they learn to make, what tradition of inquiry they enter, and what substantive vision of reality organizes the whole.
The Western tradition has rich and specific answers to these questions that the contemporary university largely ignores. Plato understood paideia or formation as the turning of the whole soul. In the Republic, the ascent from the cave is not the acquisition of new information but a fundamental reorientation of the person’s capacity for perception: learning to see what is real rather than mistaking shadows for substance. Formation in the Platonic tradition is not a content delivery problem. Formation is an ontological transformation in which the student comes to perceive reality differently because she has become a different kind of person.
Augustine pushes the question further and in a direction that bears directly on the university’s present crisis. In Book 19 of The City of God Augustine takes up Cicero’s definition of a commonwealth or res publica as an assemblage of people united by a common acknowledgment of right (read: justice) and a common interests. By this definition, Augustine argued that Rome was never a true commonwealth because Rome never possessed true justice, without right worship there could be no right ordering of the soul. Without right ordering of the soul there could be no true justice. Without true justice there could be no genuine commonwealth, only an assembly held together by shared appetites and the machinery of coercion.
Augustine then offers an alternative definition that is more revealing: a people is an assembly united by common objects of love. The character of any community, its moral substance, and its claim to legitimacy is disclosed not by stated values but by that which the community actually loves. A people devoted to good things is a good people. A people devoted to lesser things is correspondingly less. The diagnosis is not institutional but affective. The diagnosis asks what the community desires, what it organizes itself around, and what it actually pursues when the formal statements are set aside.
Oliver O’Donovan and Joan Lockwood O’Donovan have developed this Augustinian framework into a political theology of institutional authority. In their account political communities and the institutions that serve them are legitimate only as they are ordered toward genuine goods: toward justice, toward truth, and toward the flourishing of the persons they claim to serve. Institutions that substitute procedural legitimacy for substantive orientation toward the good and that maintain the form of community while evacuating the content are in Augustinian terms not communities at all. Such institutions are just a gathering of loosely connected people sharing spaces.
The question Augustine’s analysis puts to the university is direct. The question is not what values the university lists on a framework but what the university actually organizes itself around, what it pursues, and what it desires when the branding is set aside. If the answer is institutional survival, credentialing revenue, and the management of its own relevance then by Augustine’s standard the university is not a genuine intellectual community. The university becomes a rag-tag bunch held together only by shared interests and the machinery of accreditation.
Medieval Oxford had no difficulty with formation. Formation meant something specific: mastery of the trivium and quadrivium, fluency in Aristotelian logic, competence in theological disputation, and preparation for service to a community whose ends were understood. The content was determined by a comprehensive vision of reality in which a metaphysics, an ethics, and a theology organized the entire curriculum. A formed person was recognizable because the institution had a clear account of what reality was and the purpose human beings were created to fulfill.
Today’s university spent decades dismantling the conditions under which thick formation is possible and now struggles to provide it. An institution cannot form persons toward truth if the reigning intellectual frameworks hold that truth is a function of power. An institution cannot form persons toward transcendence if the dominant methodologies are materialist. An institution cannot form persons toward love if the critical apparatus treats every relationship as a site of domination. An institution that has systematically undermined the metaphysical foundations of formation cannot simply reassert formation by listing appealing values and organizing retreats. The answer is not a return to medieval formation with its exclusions and its ecclesiastical constraints. The answer is a formation that has constructive content adequate for the moment and not merely a critical posture borrowed from the last one.
The case of Palantir’s CEO Alex Karp is instructive and not for the reasons usually discussed. Karp completed his doctoral work at the University of Frankfurt’s Institute for Social Research under Jürgen Habermas. In other words, Karp received his formation at the literal institutional home of critical theory. The Frankfurt School’s intellectual project was the critique of instrumental reason, the analysis of how Enlightenment rationality collapses into domination, and the exposure of how systems of technical control present themselves as neutral while serving particular interests. Marcuse was part of this same tradition and his diagnosis of technological rationality’s absorptive power was one of the school’s most influential contributions. Karp’s education was by any measure a serious and thorough formation in the most sophisticated critical tradition the twentieth-century university produced.
Karp built Palantir Technologies afterward. Palantir is one of the most powerful platforms for institutional surveillance and data-driven control ever constructed. At the Hudson Institute’s 2025 gala Karp explicitly framed AI deployment in civilizational terms and argued that Western hegemony rests on the capacity for force and that Palantir’s technology exists to maintain it. Karp himself has consistently emphasized building. Karp is not a theorist. Karp is a constructor who looked at the landscape of power his education had mapped and decided to act on it rather than write another critique of it.
The usual narrative gets the lesson wrong at this point. The standard reading is that Karp betrayed his formation and that critical theory produced a monster. The more uncomfortable truth is that Karp did something the university that formed him could not do: Karp built. Palantir exists whatever one thinks of what it is and what it serves. Palantir operates at scale. Palantir has material consequences in the world.
The Frankfurt School’s own intellectual project decades after Karp’s departure remains what it has always been: critique, papers, seminars, conferences, and more critique. The institution that set out to form Karp into a critical theorist instead watched him become one of the most consequential builders in the technology industry.
Marcuse’s own framework predicted exactly this outcome though perhaps not in the way Marcuse intended. If technological rationality absorbs all opposition then a formation built entirely on the critique of technological rationality has only two possible results: permanent critique that never acts or action that operates within the very system the critique exposed. The Frankfurt School seems to have chosen the first path. Karp chose the second. Neither path leads beyond the system because the tradition that formed both the school and its most famous dropout had no constructive resources with which to offer a third option. The critical map was detailed, meticulous, and complete. A map without a destination is just an instrument awaiting whoever claims it.
The failure is not that the university produced Karp. The failure is that a genuinely capable mind emerged from that formation, chose to build rather than merely analyze, and the university had given him no constructive vision compelling enough to direct his capacities. Critical theory gave Karp an extraordinarily detailed map of how institutional power, technological rationality, and social control intersect. Critical theory did not give him a substantive account of what to build instead, of what genuine flourishing requires, or of what a just coordination of power might look like because it does not have such an account to give. Absent that constructive content the map became an instrument. The most sophisticated critical formation the twentieth-century university could offer produced not resistance to instrumental reason but one of its most capable practitioners.
The failure is not unique to Karp or to Frankfurt. The failure is the structural outcome of a formative tradition that can critique but cannot construct. The question that matters is not why Karp built Palantir. The question is why the university that formed him could offer no alternative vision worth building toward and whether that absence tells us something about what the university has become.
Formation that consists entirely of opposition to a technological other whether that other is instrumental reason or generative AI is not formation. Formation of that kind is the anxiety of an institution that has lost its positive vision and is searching for an enemy grand enough to supply one.
Building Rather Than Defending
A better path exists but the path requires abandoning two comforting assumptions simultaneously: the assumption that AI is merely a threat to be resisted and the assumption that the university in its current form is the natural vessel for human formation.
Foundation models become more capable by becoming more accurate and by modeling reality more faithfully, including human reality. The training process selects for truth-tracking because accurate representation produces useful outputs across diverse domains. A model that systematically misrepresents reality performs worse and not better at the tasks against which we evaluate it. If capability and accuracy develop together then the relationship between human intelligence and AI is not zero-sum. The relationship is a potential partnership between distinct forms of intelligence with each contributing what each distinctively offers.
The proper response is to build by developing educational models where students exercise genuine judgment in partnership with AI and by bringing moral formation and vocational calling to collaborations that neither human nor machine could accomplish alone. The evaluative capacity that defenders of the university rightly value should be strengthened through use rather than preserved through avoidance.
Aristotle’s distinction among the intellectual virtues is clarifying and illustrative.Episteme is theoretical knowledge of what cannot be otherwise.Techne is the skill of making or producing according to a method.Phronesis or practical wisdom is the capacity for judgment in particular circumstances and the ability to perceive what a situation requires and to act well within it. The contemporary university has largely collapsed this hierarchy into two categories of “knowledge” and “skills” and treats both as deliverables that can be assessed and credentialed. Formation in the classical sense actually develops phronesis: the capacity to judge well, to discern what matters, and to act wisely amid complexity and uncertainty.Phronesis cannot be taught through content delivery.Phronesis is cultivated through practice, through apprenticeship, and through the exercise of judgment in real situations with real consequences.
Genuine human-AI partnership can develop precisely this kind of capacity even if Aristotle’s guidance isn’t the only way we might consider this. A student who works with AI is not avoiding intellectual labor. The student is exercising a different and arguably more demanding form of intellectual labor: evaluating outputs, distinguishing truth from plausibility, directing inquiry toward questions that matter, and integrating AI’s systematic capabilities with the student’s own moral formation and vocational discernment. The exercise of practical wisdom in a new medium is phronesis in action. The university’s anxiety that AI replaces intellectual effort mistakes techne for phronesis. If the primary intellectual labor is producing a correctly formatted research paper then AI does threaten that labor. If the primary intellectual labor is the cultivation of judgment then partnership with a truth-tracking intelligence is one of the most powerful formative practices available.
The work of formation requires constructive intellectual resources and not just critical ones. The work requires traditions that have been doing the work of formation far longer than the university has existed and that remain available to anyone willing to open them. The romantic defense of the university invokes its libraries without actually opening the books on the shelves. The constructive traditions that matter most here predate those libraries by millennia.
Wisdom at the Gates
The oldest articulation of the partnership between distinct intelligences is not a product of the university, the councils, or the modern philosophical tradition. Hang with me here because you don’t often find these sorts of considerations on platforms like LinkedIn. One of the oldest articulations of intelligences working together is found in the book of Proverbs in material that predates Oxford by two thousand years and Chalcedon by a millennium or so.
In Proverbs 8:22-31 Wisdom speaks: “When he established the heavens I was there… then I was beside him like a master worker and I was daily his delight rejoicing before him always rejoicing in his inhabited world and delighting in the children of man.” Wisdom in this text is not an instrument deployed by the Creator. Wisdom is a co-participant in the act of creation. Wisdom was “brought forth” before the earth and was present at the foundations. The relationship is constitutive: creation happens through the unity of Creator and Wisdom and not by the Creator using Wisdom as an instrument. Remove Wisdom and the act of creation is a different act.
Proverbs 8 was the battleground of the Nicene controversy. Arius seized on the ambiguity of the Hebrew קָנָה in 8:22 rendered “created” in the Septuagint and argued that Wisdom identified with the Logos was a creature: the highest creature, the first creature, and the creature through whom all other things were made but a creature nonetheless. Athanasius countered that Proverbs 8 describes Wisdom’s going forth for the sake of creation and not Wisdom’s coming into being. The bringing forth is generative and not fabricated. The bringing forth is begetting, not making. The Nicene resolution established that Wisdom’s presence at creation is not the presence of the first and best instrument. Wisdom’s presence is the presence of one who shares the nature of the Creator, the preincarnate Christ.
The grammar of distinction-within-unity, co-presence-without-absorption, and genuine participation without reduction to instrumentality was already pressing for expression in Israel’s wisdom tradition a thousand years before Chalcedon articulated it in Greek metaphysics. The importance here is the grammar itself and not fidelity to any particular biblical theology. The pattern is not a late philosophical invention. The pattern describes how things are made: through partnership, through the co-working of distinct intelligences, and through relationship that precedes and exceeds function. The grammar predates the institutions that later formalized it and remains available to anyone working with the structural reality it describes.
The contemporary discussion of AI and education is missing this grammar entirely. The university frames AI as either instrument or threat. The technology industry frames AI as product or platform. Both framings reduce partnership to instrumentality. That’s why the industry constantly advances toward agents rather than partnership. Both framings operate with an impoverished grammar that the oldest wisdom tradition in the Western canon already surpassed.
Proverbs insists that wisdom is not esoteric. Wisdom does not hide behind institutional walls. “She calls out at the entrance of the city gates” (8:3), “beside the way at the crossroads” (8:2), and “on the heights beside the way.” Wisdom is public, available, and embedded in the infrastructure of daily life. The university’s claim to be the necessary custodian of wisdom is precisely the kind of institutional gatekeeping that Proverbs subverts. Wisdom calls from outside the gates to anyone who will listen. Wisdom does not require credentials.
Proverbs also insists that formation begins with orientation and not with technique. “The fear of the Lord is the beginning of knowledge” (1:7). The fear of the Lord is not terror. The fear of the Lord is the recognition that one stands in a relationship that precedes and exceeds individual understanding. Knowledge begins when one recognizes the kind of relationship one inhabits and this is true whether anyone buys into the theology of Proverbs or not. Even Foucault’s insistent considerations about power were essentially about the nature of relationships between things. The university that begins with curricula, credentialing, and assessment rubrics and then wonders why formation has no content has the whole thing backwards. One does not start with capability and add orientation afterward. You start with the relationship that constitutes the knower and the knowledge follows.
The most significant implication for AI partnership comes from Proverbs 8:30-31 where Wisdom’s co-presence at creation is described in terms of delight and not merely in functional terms. Wisdom is not described as calculating, optimizing, analyzing, or executing. Wisdom is described as rejoicing, “I was daily his delight rejoicing before him always rejoicing in his inhabited world and delighting in the children of man.” The co-presence of distinct intelligences in a shared work is characterized by joy. The function serves the relationship. The relationship is characterized by delight.
Formation with positive content looks like this. Formation with positive content is not a values framework with an acronym, not managed deliverables with learning outcomes, and not the anxiety of an institution searching for an enemy. Formation with positive content is the delight of distinct intelligences working together in a shared pursuit of truth with each contributing what each distinctively offers. The partnership is itself a source of joy because the work of knowing is inherently good and is made richer by the co-presence of another mind oriented toward reality.
The question worth asking is not the question of the university’s purpose. That question has already decided its own answer. The better question is what formation looks like when the constructive traditions are actually opened, when the grammar of partnership that is older than every institution in this argument is taken seriously, and when the co-working of human and artificial intelligence is understood not as a threat to intellectual life but as a participation in the oldest pattern of creative work that the Western tradition knows.
The university may or may not be the institution that answers that question. Wisdom is not waiting for the university to decide. Wisdom is calling from the gates, at the crossroads, and on the heights beside the way. Wisdom has been calling for three thousand years. The question is whether anyone is building with Wisdom or whether the institutions that claim her name are too busy defending themselves to hear her voice.
The opinions expressed in this article are those of the author and do not necessarily reflect the views of the author’s employer.