Taking a Student Success Approach to Higher Education AI Policy

Working in a college or university right now might feel like “AI policy” is taking up all the oxygen. Many, if not most, campuses have launched AI task forces charged with developing policies, and they are understandably struggling to accommodate multiple criteria that may be in tension. For example, how can an AI policy set firm boundaries, nudge or shape behaviors that align with the educational mission, and allow for the flexibility required by an institution’s traditions of faculty autonomy, shared governance, and disciplinary integrity?

It would not be surprising if discussions about AI policy are raising more fundamental questions about mission and strategy. In two earlier articles, I argued for the need for more institutional-level leadership responding to AI and proposed a guiding star for AI strategy. I made the case that waiting to see what interesting experiments with AI bubble up is too slow a process for the speed of AI adoption and not student-centered enough. Institutions need to respond to the defining leadership moment AI is presenting, and they need to take approaches to AI that prioritize student success for every learner.

In this article, I want to look more at the possible AI policies that might result from institutional leadership that begins with a student-success framing. Confronting tensions between exploration, integrity, assessment, and autonomy is critical because students experience the consequences of muddled AI policy in their courses. For example, in Student Research Into How Students and Faculty Use AI, which was carried out by Every Learner Everywhere student interns, one finding was how alert students are to inconsistent AI policies. They are asking valid questions about why one class absolutely forbids its use and another actively embeds it into assignments and assessments. Similarly, they notice that one student may get very little experience with or guidance on AI while a peer in another major gets lots of hands-on practice. Some individual students in the survey said that because they were unclear on policy, they avoided AI use, perhaps unnecessarily, out of fear of being falsely accused of cheating.

A 2026 survey from Lumina Foundation shows something similar. Over half of students in it say at least some of their courses have unclear policies. Incidentally, those policies often have little influence on students’ actual behavior. Among students at schools that discourage AI use, half still do so at least weekly.

This is not a situation that tends toward supporting student success for every learner. Nor does it tend toward supporting other goals an institution may have, such as innovation or operational efficiency. As long as people are having to guess what AI use is acceptable, AI policy is not keeping up.

Considerations for AI policy

A student success framing gives institutions a practical way to test AI policy choices. Policies should make acceptable use clear, protect students from avoidable risks, and leave room for faculty and staff to apply AI in ways that fit disciplinary and institutional goals. These five considerations can help.

Accessibility, bias, and privacy

Policies should set guardrails for data governance, accessibility standards, and bias mitigation for any AI used in teaching or operations. The goal of the policy should be that it works to further the success of every learner. As I discussed in my first article, higher education’s fragmented approach to digital learning in general over the last 20 years left too many learners behind.

Mission and values alignment

Policies should explicitly state how AI supports institutional commitments to academic integrity and student success, avoiding generic or purely prohibitive language. The goal is not simply to define what is and is not permitted but to allow innovators to see where AI use can potentially help them improve teaching and learning.

Audience-specific clarity

Colleges and universities need guidance that is consistent but tailored to account for several facets of the work they do, such as institutional policy, faculty syllabi language options, student-facing explanations, and staff use guidelines.

Transparency and disclosure

Policies should clarify expectations for when AI use must be cited or disclosed by students, faculty, and staff. Again, while this discussion understandably originates with concerns about academic and professional integrity, I encourage a student-success framing. The goal should be to help students understand how transparency and disclosure relate to developing communication, analysis, and other academic skills.

Flexibility and iteration

Policies should be framed as living documents. The institution should establish review cycles and feedback channels so AI policies can evolve with the technology and as evidence accumulates.

The role of faculty and staff in AI policy

Implementing an AI policy can easily become a “check the box” exercise if institutions expect faculty and staff to do so with no time, training, or support. If an institution wants responsible AI use in teaching, it must resource faculty and staff to learn, test, and iterate together. Consider these roles that individual stakeholders can play.

Champions and innovators

To begin developing an AI policy, and to test drafts against real-world conditions, begin by finding the people on campus who are eagerly exploring AI tools. It’s a good bet that they are keeping notes about what is and isn’t working.

Culture carriers

It’s important to think about how responsible AI use will be modeled and normalized in operations, documentation, communication, teaching, and learning across the institution. Those early adopters and champions are also likely to be effective ambassadors for your AI policy later.

Translators and integrators

Clarify how staff in centers for teaching and learning, academic technology, advising, and institutional research will translate high-level AI vision into workflows, training, and support that faculty and students actually experience. Emerging practice from faculty-led AI groups shows that communities of practice, templates, and shared language (e.g., recommended syllabus statements) reduce burden and inconsistency.

Early-warning sensors

Points of friction, such as new accessibility barriers for students or new workload issues for faculty and staff, can emerge in unanticipated places. It might be in the registrar’s office or the bookstore, for example, where a policy gets tangled up. The Every Learner website recently featured the story of how the women’s volleyball coach on one campus shared the crucial insight for improving how digital learning was being deployed in academic courses. Institutions should work to ensure that the individual faculty and staff who address the points of friction students experience should be incorporated into the feedback channels as the AI policy is reviewed and revised.

AI policy as an expression of institutional leadership

In my earlier two articles I outlined how the response to this “AI moment” goes to the core of mission, strategy, and governance for institutions of higher education. A lot more goes into that response than an AI policy alone, but the AI policy is a key expression of institutional leadership. It will inform student learning experiences in practical ways on a daily basis. To ensure those experiences are coherent rather than fragmented, colleges and universities should keep student success for every learner as their guiding principle in developing an AI policy.