AI Policy Lab

A faculty workshop for reading, analyzing, and designing AI use policies at course, program, and department levels.

Length
Three hours, as one session or two 90-minute sessions. CWSP facilitates.
For
Individual faculty, departments, curriculum committees, department heads, and associate deans. Works best with 8 to 20 participants, from one department or a mixed group.
You leave with
A policy draft at the course, program, or department level, grounded in your own disciplinary reasoning.

What the AI Policy Lab does

Most faculty make AI policy decisions alone. They work from a draft syllabus statement, a department memo, or institutional guidance written under pressure, which might not reflect what anyone believes. The AI Policy Lab gives that work a structure: an analytic method, a comparative corpus, and colleagues from across the university.

The workshop uses the AI Governance Census of US Colleges and Universities (opens in a new window) (Cole, 2026, version 0.5, doi:10.5281/zenodo.23887259 (opens in a new window)) as its primary texts: a stratified census of 334 institutions drawn from the 2025 Carnegie Classifications, of which 90 have been searched so far, yielding 1,125 document rows, 724 of them verified. Eighty of the 90 are research universities, so the figures describe R1 governance until the remaining strata are in.

Twenty-one of the searched institutions are land-grant universities, and the workshop draws its comparative cases from them: fourteen of the twenty-one hold contradictions between their own documents, eighteen have announced bodies or policies with no published text, and eleven have integrity codes which do not name AI.

Participants read those documents for the governance problems they reveal. What did institutions say when they did not yet know what to say? And what do those moments expose about institutional values, authority, and the limits of policy? And where two current documents on the same campus state opposed rules, which one governs?

Participants leave with a policy draft, at whatever level is most urgent for them, grounded in their own disciplinary commitments and written to hold up through the next product cycle.

Workshop materials

The documents below work as a set, and analyses and drafts written in the workbook go to CWSP, apart from the census.

The analytical framework

The workshop is built on two tools. The first is a four-step analytical procedure: identify the document's genre, locate the glitch (the moment when the document strains against itself), read what the glitch reveals about institutional values, and name the abstraction (the general claim about AI governance or authorship that the document makes legible). The second is a three-level literacy framework (Cole 2026) that asks, of any policy document, what it demands from readers at each of three levels: conceptual description, structural analysis, and operative decision. A policy written at the first level alone names the tools. One written at all three states what the institution analyzed and what it decided, with reasons.

Together, these tools make policy reading a teachable skill. Participants take it back to their departments and apply it to new documents as they appear.

Workshop structure

The workshop runs in four modules. Module 1 introduces the policy corpus and shows why AI governance documents reward careful reading for what they reveal under pressure. Module 2 applies the four-step procedure to cases from the census's record of contradictions (opens in a new window), adds the question the procedure lacked (whether a conflict is a contradiction between two current documents or a supersession, where one has replaced the other), and reads NC State's own record against its land-grant peers: eleven documents, seven of them published and verified, three promised with no text yet, and two OIT pages which disagree on the data the approved tools are cleared to hold. In Module 3, a facilitated design activity, participants draft or revise policy language at the level most relevant to them, using adaptable templates calibrated to the five principles of functional AI policy: scope terms by function, date the policy, name the revision trigger, separate durable principles from contingent prohibitions, and explain the reasoning. Participants also work through one institutional decision at full depth, the procurement of AI detection software, on the framework page's decision tool. Module 4 sets up a review cycle, because the census finds about five percent of governance pages moved or changed within a single week of verification, and 134 announced policies or bodies with no published document at 59 institutions, so a policy drafted this term will meet campus rules which arrive after it. Participants leave with a term-by-term check of their own policy against the census record and their own document links.

AI-writing detectors

Departments drafting academic integrity language often ask where CWSP stands on AI-writing detectors. The program does not use them, and the decision and its reasons are on the record, written to the five principles the workshop teaches: it defines the tool by what it does, carries a date, names the condition for its revision, separates the durable principle from the contingent rule, and gives its reasons. Penn State's academic integrity office states the same position in one sentence, which the workbook's course template quotes.

Relationship to AI Literacies and WOLF

The AI Policy Lab stands alone, and it also fits within the wider set of WOLF programs. Faculty who have worked through the AI Literacies extension bring a vocabulary for what AI tools do to writing instruction. The Policy Lab puts that vocabulary to work at the institutional level. For faculty new to AI Literacies, the Policy Lab is a good place to start, particularly when their immediate concern is policy language and assignment design comes later.

The analytical procedure at the center of the Policy Lab (read for the genre, locate the glitch, read the reveal, name the abstraction) is also the core literacy practice of the AI Literacies extension. Faculty who complete both work with the same skills in each, applied to policy in one and to teaching in the other.

Related talks

All research and presentations

Request a workshop

To request an AI Policy Lab workshop for your department or curriculum committee, use the intake form (opens in a new window) or write to Kirsti Cole at CWSP.