Key points:
- Governance in the age of AI is ultimately about preserving institutional agency
- AI as a scaffold for learning: From access to judgment
- Recentering discovery as the core of learning through AI
- For more news on AI and governance, visit eCN’s AI in Education hub
Higher education is rapidly developing AI governance frameworks through the creation/modification of policies, establishing compliance structures, conducting procurement reviews, and developing acceptable use guidelines. These efforts are necessary, but they are also increasingly insufficient because they are built around an assumption that governance is primarily about controlling technology and managing risk.
While this may have been adequate in the past, AI fundamentally changes the equation. Institutions are no longer simply governing software platforms or isolated tools, but systems capable of continuously shaping decisions, behaviors, incentives, and institutional outcomes at scale. Existing governance models in higher education evolved over time around relatively stable systems, focusing on episodic human decision making, and in environments in which institutional change occurred slowly enough for layered review, procedural oversight, and committee deliberation to function effectively.
AI fundamentally alters these conditions with intelligence becoming continuously available, adaptive, interactive, personalized, and scalable across virtually every institutional function. In many cases, due to use of third-party software/platforms, institutions may not even be aware of the extent of AI use, although they still bear responsibility for its use. The challenge is not that institutions are moving too quickly towards AI adoption, but that higher education is approaching AI primarily as a technology implementation problem rather than as a structural transformation of learning, and hence how institutional value is defined and operationalized.
The central governance challenge in the age of AI is therefore no longer simply governance of compliance, but rather that of governance of purpose. Governance of compliance asks whether systems satisfy policy, regulation, privacy, cybersecurity, procurement, and procedural requirements. These questions remain essential, and institutions that fail to address them expose themselves to significant operational, legal, financial, and reputational risks. But governance of purpose asks a fundamentally distinct set of questions:
- Are institutional systems producing outcomes consistent with mission?
- What institutional behaviors and metrics of success are being normalized over time?
- Are optimization objectives beginning to value efficiencies entirely over educational purpose?
- Are patterns of decision-making reinforcing or weakening educational goals?
- Who is accountable when technically successful systems produce institutionally problematic outcomes?
AI presents a fundamentally different challenge from that previously faced by higher education because it does not simply represent the digitization of existing processes, but alters the condition under which knowledge is accessed, expertise is developed, and learning occurs. Intelligence is becoming continuously available, interactive, personalized, adaptive, and most importantly, scalable. The implications extend far beyond classrooms, affecting aspects such as admissions, advising, financial aid, hiring, curriculum development, student support, research administration, operational planning, workforce preparation, and institutional strategy itself.
A university can satisfy every current governance checklist and still fail institutionally. An admissions system may optimize enrollment yield while gradually undermining access and socioeconomic mobility. A student success platform may improve retention metrics while discouraging intellectual risk taking or narrowing pathways of exploration. Financial aid optimization systems may maximize net tuition revenue while shifting institutional priorities away from mission-centered student support. AI-generated curricula and credential completion pathways may optimize graduation rates while weakening intellectual development and preparation for the workplace. In each case, every technical control may function exactly as designed, with failure due not to violation of policy, regulation, or specification but rather because governance remains focused on operational compliance while insufficient attention is paid to whether the aggregate direction of institutional behavior remains aligned with institutional purpose. This is the governance challenge that higher education has not yet fully confronted.
Most current discussions of AI governance in higher education focus appropriately on critical areas such as cybersecurity, infrastructure, data governance, model validation, procurement and vendor integration, compliance, privacy, and risk management. These issues are essential, and institutions that fail to address them expose themselves to the significant operational, financial, legal, and reputational risks. Yet these do not completely or even adequately address governance. Traditional governance structures functioned reasonably well in environments where operational cycles move slowly enough for institutions to evaluate consequences before decisions propagated broadly across the institution. Governance occurred through layered review, committee deliberation, procedural oversight, and distributed authority. AI changes that environment completely.
The central question before higher education is not whether AI should be adopted. The more consequential question is whether institutions are willing to rethink long-standing assumptions about instruction, assessment, governance, and institutional relevance in an environment where intelligence is no longer scarce. That distinction matters because governance frameworks built primarily around control and procedural oversight are increasingly mismatched to systems operating continuously, adaptively, and at scale. When AI-based systems operate continuously across institutional functions, speed itself becomes part of the challenge because flawed assumptions, poorly aligned incentives, incompletely defined objectives, or problematic optimization priorities can now scale rapidly across thousands of decisions, interactions, and recommendations, simultaneously. Previously, a problematic policy might affect dozens of people before concerns emerged. The same flawed logic embedded within adaptive AI-supported systems can now shape admissions decisions, advising patterns, financial aid recommendations, student interventions, and resource allocation decisions continuously, and at scale. This is why AI governance cannot be reduced to policy creation alone.
Policies matter, but policies by themselves cannot govern systems operating beyond the pace and scale of traditional institutional oversight mechanisms. Governance in the age of AI must increasingly focus not only on whether systems comply with requirements and regulations, but also on who has the authority to question objectives, challenge assumptions, evaluate outcomes, and remain accountable for the institutional behaviors these systems produce over time. That distinction represents a profound shift. One of the emerging risks in higher education is that AI governance becomes overly operationalized, concentrated primarily within technology, compliance, procurement, legal, cybersecurity, or risk management structures, and is deliberated upon, developed, and implemented by people with IT and technological backgrounds without any direct academic experience. These experts play indispensable roles, but governance becomes incomplete when it is confined primarily to operational and IT domains neglecting those most personally responsible for educational purposes, intellectual development, and academic integrity.
Faculty cannot be peripheral to AI governance conversations because universities derive legitimacy from their direct interaction with students and their explicit responsibility to cultivate knowledge, ensure learning and assure preparation for professional and societal success. When AI systems increasingly influence learning pathways, assessment, engagement, decision making, and institutional priorities, governance must include those capable of asking not only whether the systems function effectively, but whether they reinforce or distort the very purposes institutions exist to serve. That challenge becomes even more significant as higher education increasingly adopts AI not simply as administrative and academic support, but as part of the architecture of learning itself.
Recent discussions around AI in higher education have often framed the technology through the lenses of access, acceleration, productivity, or automation. While these dimensions matter, they risk understating the deeper transformation underway in re-envisioning how learners engage with information, develop judgment, explore complexity, and receive feedback. It is creating the possibility of forms of iterative, personalized, and continuously responsive learning that were historically difficult to scale. This creates extraordinary opportunities for institutions willing to rethink educational design rather than simply digitize existing practice, but it also raises governance stakes substantially because institutions are no longer simply governing technologies layered onto existing environments but are increasingly governing systems that influence the formation of expertise itself. That distinction carries significant implications for institutional responsibility.
For decades, higher education has operated within structures shaped by controlled scarcity of access to knowledge, expertise, networks, and credentials. AI fundamentally alters that structure. As intelligence becomes increasingly abundant and continuously available, institutional value shifts away from access to ability, and from teaching to learning through designed experiences. Governance must evolve accordingly. AI-enabled systems continuously reinforce specific incentives, behaviors, priorities, and definitions of success. Over time, these optimization pressures can reshape institutional behavior even in the absence of explicit policy change. The governance challenge is therefore no longer simply governing institutional activity and interactions, but governing what institutions are systematically being optimized to become. The risk is therefore not just flawed AI implementation, but the deeper risk of institutions gradually surrendering institutional direction to optimization algorithms governed primarily by operational efficiency, measurable outputs, short term performance metrics, and fragmented incentives. Governance in the age of AI is therefore ultimately about preserving institutional agency. This requires moving beyond static governance framework towards adaptive governance architectures.
Adaptive governance does not mean abandoning oversight or lowering standards but rather requires recognizing that governance systems themselves must become more continuous, iterative, interdisciplinary, and outcomes aware. Institutions must develop mechanisms capable of evaluating not only isolated decisions, but cumulative institutional effects, especially those that impact the process of learning itself.
This includes asking difficult questions such as:
- What patterns emerge across thousands of AI supported decisions?
- What institutional behaviors are being reinforced?
- What forms of student engagement, faculty work, scholarship, and learning are being privileged or marginalized?
- Are institutional incentives drifting away from mission?
- Are efficiency metrics beginning to dominate educational values?
- Are institutions optimizing for measurable outputs while weakening less quantifiable dimensions of intellectual development?
These are not hypothetical concerns. They are precisely the type of structural questions governance exists to address and currently fail in large measure. This is why AI governance cannot remain narrowly centered on compliance, cybersecurity, procurement, or technical oversight alone. Those functions matter, but they address only part of the challenge. The deeper challenge is whether institutions of higher education remain capable of exercising intentional institutional direction in environments increasingly shaped by continuously operating intelligent systems.
Governance in the age of AI is therefore no longer simply about controlling technology. It is about preserving institutional purpose, academic integrity, public legitimacy, and the capacity for institutional self-determination before optimization systems, operational pressures, and fragmented incentives begin governing institutions in their place.
The institutions that will lead in the coming decade will not be those that deploy AI most aggressively or restrict it most cautiously. They will be the institutions that recognize that AI has fundamentally changed the operational architecture of higher education itself, understand that governance is no longer merely about controlling people and systems, and respond by redesigning governance around mission accountability, institutional purpose, and continuously evaluated outcomes. In the future, institutional failure will not occur simply because systems lacked safeguards, cybersecurity protections, or regulatory compliance.
In the age of operationalized intelligence, the defining governance challenge for higher education is no longer simply how institutions use AI. It is whether institutions remain capable of governing what AI systems are continuously shaping them to become. Failure will increasingly occur because institutions allowed AI systems to optimize efficiency, scale, revenue, retention, or operational convenience without adequately governing whether those outcomes advanced the deeper purposes universities are meant to serve.
Vistasp M. Karbhari is a Professor in the Departments of Civil Engineering, and Mechanical & Aerospace Engineering at the University of Texas at Arlington, where he served as President from 2013-2020.