University of Wisconsin–Madison

Keeping pace with student use of AI

Walk into any classroom, lab, or clinical skills session across our health professions programs, and you will find learners using generative AI to study, summarize, draft, and explore. These tools can produce text, images, or code from prompts, and the technology has arrived faster than the rulebook. So, this month we ask: What does responsible AI use look like in our coursework and assessments, and how do we make expectations clear to every learner in every program?

A 2026 cross-sectional analysis in The Clinical Teacher reviewed AI‑related policies across 199 accredited U.S. medical schools. Two in five schools had no AI policy at all, and many existing policies originated at the larger affiliated university rather than the medical school. Those policies rarely address clinical documentation, patient privacy, or the distinct demands of professional formation. The authors describe this “policy gap” as a risk for unclear expectations, inconsistent standards, and academic integrity disputes.

A 2026 article in International Medical Education argues that sustainable solutions lie not in blanket restriction but in principled integration: redesigned assessments, AI literacy paired with professionalism, and coherent governance. In other words, the question is not whether students will use AI — but how we teach them to use it well.

Our peers are moving. At Stanford Medicine, a generative AI policy now covers both MD and physician associate (PA) students, pairing integrity expectations with instruction on responsible AI use. At Oregon Health & Science University, a graduate medical education AI policy requires programs to state expectations in every syllabus, asks trainees to disclose and attribute AI assistance, and treats undisclosed use as an integrity violation. Nationally, the AAMC’s Principles for the Responsible Use of AI in and for Medical Education offer a shared framework for doing this well.

We are not starting from zero. UW–Madison’s Center for Teaching, Learning and Mentoring offers guidance and workshops to help instructors set clear AI expectations — whatever position they take on its use. The university provides vetted tools and guidance on the many types of data with which they can be used. On the SMPH intranet, the AI‑powered Medicine resource collection gathers policy templates, sample syllabus language, and teaching examples — including assessment approaches — for faculty and staff. Through the Wisconsin Research, Innovation and Scholarly Excellence initiative (RISE‑AI), our school is also helping lead the university’s investment in AI research and training for practitioners.

What we do not yet have — what few schools have — is a shared, cross‑program understanding of the rules. That conversation belongs to all of us: MD, PA, DPT, genetic counseling, and MPH alike. In the months ahead, Academic Affairs will begin convening program leaders and interested faculty to explore a shared approach to AI use in coursework and assessment.

Dig Deeper

Have a smart question about smart technology? Send it our way — a future edition of CurrentAffairs may take it on. Email blsilver@wisc.edu with ideas.