By Joseph Beam, Director of Marketing and Communications, AACRAO, Live from #AACRAO2026
I arrived early for the “Ctrl+F for Compliance: AI and Enrollment Reporting” session hosted by Dr. Rhonda Kitch and Gena Boling on Monday morning at the 111th AACRAO Annual Meeting, and within minutes, every seat was filled, with latecomers lining the walls and sitting on the floor.
The team from Cornell University opened with a perspective that reframed artificial intelligence not as a recent disruption, but as part of a much longer institutional story. Cornell’s engagement with AI dates back to 1958, with the development of the Perceptron, the first artificial intelligence purported to have an “original thought.” It served as a reminder that while today’s tools may feel novel, the intellectual groundwork has been unfolding for decades. What feels new, perhaps, is not the technology itself, but the immediacy of its application across administrative functions like compliance and others.
What struck me was how intentionally Cornell has approached this challenge.
As a complex, decentralized institution with multiple schools, systems, and layers of authority, Cornell could easily have fallen into fragmented experimentation. Instead, the creation of its AI Innovation Lab in 2024 appears to have provided both structure and momentum. With more than 30 projects already underway, the lab is not simply testing ideas but actively integrating AI into operational workflows. The examples shared during the session were both practical and ambitious: using AI to identify error patterns in student enrollment records, uncover previously undetected data anomalies, and distill dense regulatory guidance into more accessible, actionable insights specific to users’ needs. Even better, the session was not a showcase of tools so much as a case for process (a personal favorite of mine).
One of the clearest takeaways was the importance of starting with the problem, not the technology. The presenters emphasized a disciplined approach to identifying gaps, asking where inefficiencies exist, and determining whether AI is truly the right solution. This may sound straightforward, but in practice it requires a kind of institutional restraint. There is a temptation, especially now, to lead with capability rather than need. Cornell’s approach offers a useful counterpoint: define the problem, evaluate the fit, and only then consider implementation.
Equally compelling was the discussion of human-centered design and ethics.
AI, as the presenters framed it, is only as effective as the values embedded within it. This was particularly relevant in the context of enrollment compliance, where decisions carry real consequences for students and institutions alike. Ensuring transparency, maintaining accountability, and aligning AI use with institutional ethics are not abstract ideals but operational necessities built into the process, and reinforced by a dedicated team behind operationalizing it. The session reinforced that ethical considerations must be built into the process from the beginning, not layered on after deployment.
Another theme that resonated strongly was the importance of documentation.
It is easy to think of AI as a forward-looking investment, but the presenters made a persuasive case that its success depends heavily on how well institutions understand and document their past and present (something we know a bit about regarding succession planning). Policies, procedures, and institutional knowledge must be clearly articulated if AI systems are to function effectively. Without that foundation, even the most sophisticated tools risk producing incomplete or misleading results. In this sense, documentation is not just administrative housekeeping; it is infrastructure.
The session also delivered on its learning outcomes in ways that felt grounded and actionable. It translated technical concepts into language that could resonate across institutional contexts, making it easier to imagine how similar approaches might be adapted elsewhere. There was no suggestion of a one-size-fits-all model, but rather an invitation to think strategically about environment, readiness, process, and long-term sustainability.
Walking out of the session, I found myself less focused on AI as a tool and more on AI as a catalyst. Not a shortcut, not a replacement, but a mechanism that, when thoughtfully applied, can sharpen institutional practice.
If there was a unifying message, it was this: the future of AI in higher education will belong to institutions that are willing to do the foundational work. Define problems clearly, document processes thoroughly, and align innovation with values.
The technology may be evolving rapidly, but the principles guiding its use remain firmly within our control.
share