All the Statements! A Crowdsourced Approach to Navigating AI Policy in Classrooms

In January 2023, many of us in higher education were in the same position: the semester was beginning, generative artificial intelligence (Gen AI) had abruptly entered public conversation, and faculty needed to say something about it in their courses. As Director of Faculty Development and Innovation at College Unbound, I knew I had to help faculty think about what this new technology might mean for their classrooms. At the same time, I was preparing to teach the course AI and Education and needed to determine my own policy. I found myself facing two related questions: What was my policy going to be for student use of Gen AI? And what, exactly, was I going to say to faculty who were asking versions of that same question for their own courses?

The problem was immediate. ChatGPT had been publicly available for only a matter of weeks. Faculty were trying it, worrying about it, dismissing it, imagining possibilities for it, and often doing all of those things at once. But the semester does not wait for clarity. Syllabi still had to be finalized. Courses still had to begin. Students were about to encounter assignments and expectations in classrooms where institutional guidance was absent, incomplete, or still being invented.

My response was simple: if people needed examples, why not start collecting them? I created a Google Form to gather basic details and a Google document to share submissions. It had the basics: What is the policy? What course is it for? What institution is it from? Who is sharing it? I included my own policy as the first entry and began circulating the collection across social media, listservs, and colleagues. By the end of January 2023, the repository had about fifteen contributions.

What emerged was not a solution to Gen AI policy. It was a small piece of shared infrastructure for faculty practice: an openly shared, low-friction resource where educators could contribute policy language, compare approaches, and adapt what they found for their own courses. In that sense, the Syllabi Policies for Generative AI Repository morphed a crowdsourced document into an open educational practice through which faculty could make course expectations more explicit, learn from one another’s provisional thinking, and begin to translate uncertainty into something students could actually read.

Open Educational Practice in Real Time

The repository functioned as a pliable resource educators could use, adapt, and build from without reinventing the wheel. Faculty were consulting and contributing to the collection, situating their own policies in relation to others, and making their thinking visible to a wider community.

Whenever I opened the document, other people were often already there. Emails started to arrive asking whether people could link to it, share it, use it for research, or request edit access. Contributions came from different institutions and from beyond the United States. Such signals indicated it was becoming a shared infrastructure for faculty trying to make course expectations legible in a moment of rapid change.

That openness depended on low friction. The entry point remained simple: teach a course and be willing to share a policy. People were contributing language that was provisional, imperfect, and still in formation. Contribution itself became a form of learning. To submit a policy was to articulate a stance on authorship, trust, responsibility, and student use of Gen AI in public, alongside the stances of others.

Amongst the queries that came with it, people began asking whether they could substantially adopt policies they found in the collection. I had assumed the answer was self-evident, but I had not made that explicit. In response, I added options for contributors to specify rights and licensing, including Creative Commons licenses. That change clarified reuse and strengthened faculty agency. Readers were not only invited to browse examples; they were given clearer permission to responsibly reuse, adapt, and contextualize policy language for their own courses.

Policy Language as Pedagogy

Early on for me, it became clear that these were never just policies about Gen AI. Each entry offered a small window into how a faculty member understood teaching, learning, authorship, trust, and responsibility. Every statement said something about what mattered in that course, what students were expected to do, and how an instructor understood their role in relation to student work.

The collection provided wording to borrow but it also gave faculty examples to think with and think against. Seeing restrictive, exploratory, and strongly critical policies side by side made visible a range of dispositions toward Gen AI, but also toward teaching more broadly.

Syllabi are sites of power, even when we do not name them that way. They are often framed as contracts or classroom law, and over time, many of us have filled them with increasingly dense policy language. The result can be a document full of rules but short on clarity. The repository did not solve that broader syllabus problem, but it did create a way for faculty to see how policy language shapes the relationship between instructor, student, and course.

More importantly, it helped make course expectations visible. The immediate question for many faculty was: what should my Gen AI policy be? But the more useful questions ran deeper. What does my policy communicate about my teaching? What does it suggest about authorship, assessment, or student responsibility? What kind of classroom relationship does it assume? In that sense, the repository supported more than policy drafting. It supported expectation setting.

What the Collection Revealed

Curating the collection relied mainly on whether the essential details were included, legible, and represented a real course. A handful of entries were excluded because they were hypothetical or generated only as proof of concept. Beyond that, I did not want the collection filtered according to my own preferences. Its value depended on showing what people were actually doing.

As the repository grew, patterns began to emerge. I was not building the collection as a formal research study, but after reading and revisiting hundreds of policies over time, certain tendencies became hard to miss.

More than half of the collection of policies frames AI in a conditional manner acknowledging where it is and isn’t acceptable to use while about a third of the collection fits somewhere in the restrictive to prohibitive category. These policies made clear that Gen AI use was not allowed, sometimes under any circumstances. An interesting mixture of policies emphasized the environmental, ethical, intellectual, or copyright-related problems associated with the tools, and some moved beyond caution into clear disapproval. The rest of the policies prove more exploratory and aimed to open up conversation about what they are and how they might use them.

Each policy represented a different attempt to make sense of a rapidly changing teaching environment. Each said something about how a faculty member understood the different elements that Gen AI challenged or questioned in the classroom. In that sense, the repository served a kind of microcosm of faculty response in real time.

Staying Useful Over Time

One notable aspect of the repository is not just that it was used early on, but that it continues to be used. Contributions have continued steadily, and the collection now includes over 200 entries. Just as telling, it is rare for more than a few weeks to pass without someone mentioning it in an email, at a conference, or in a professional conversation. Faculty continue to use it, share it, and return to it as they revisit their own policy language.

There are other signs of reach. Centers for teaching and learning have linked to it as a resource. Journalists have asked about it, including an interview with CNN.com. At one point, Tatiana Torres Zapata translated the collection into Spanish so it could be more accessible to a broader audience. That moment captured the power of open education as the resource moved from being accessed to being adapted and extended for another context.

At the same time, usefulness requires ongoing work. Reviewing and moving submissions into the shared collection remains manual because entries need to be checked for clarity and completeness. Perhaps, unsurprisingly, AI has not been able to help with that. As the repository grew, other adjustments became necessary. The collection moved from a document to a spreadsheet to make it easier to sort and navigate. Disciplinary categories, licensing options, and policy updates were added. Periodic reminders are also necessary, especially at the start of semesters, to encourage new contributions and revisions.

Even with this activity, the repository remains incomplete. The collection includes over 200 contributions from community colleges to R1 institutions. Public universities represent the largest contributors (80 policies). Overall though, it is still small relative to the number of courses taught each semester. Some disciplines are strongly represented such as Writing and Composition with over 50 policies while the second largest discipline (Business and Economics) has only 20 policies at the time of this writing. Not all contributors return to revise their policies. These are not failures so much as conditions of open, crowdsourced work. Openness creates the possibility of shared infrastructure, but it does not ensure full participation or continual renewal.

What This Case Offers

This repository did not produce a single best Gen AI policy, and it did not resolve the variation across courses, disciplines, or institutions. But it did offer a shared infrastructure for faculty practice at a time when institutional guidance was still emerging.

That is the transferable lesson of this case. Faculty were not only using the repository; they were learning through contributing to it. Writing and sharing a policy required them to articulate their own stance on authorship, responsibility, and student use of Gen AI in relation to others. At the same time, the ability to reuse and adapt others’ policies, supported by clearer licensing, reduced duplicated effort, and made it easier to translate ideas into course-level expectations.

The repository also highlights a practical pedagogical contribution: helping faculty make expectations visible. It did more than provide language. It supported clearer communication with students about acceptable use, disclosure, and responsibility. That function is likely to remain relevant beyond Gen AI, as instructors continue to revisit how policies shape understanding in their courses.

This project was intentionally simple. It relied on readily available tools and voluntary participation. That simplicity made it possible to build quickly and adapt over time. It also meant the repository would remain partial, dependent on ongoing contribution, and uneven in its representation. Those conditions are typical characteristics of many open, crowdsourced efforts.

What this case suggests is that such efforts can complement institutional efforts and policy development. When educators face emerging challenges before formal guidance exists, an open, adaptable, faculty-facing resource can help them make expectations explicit, learn from one another, and respond with greater clarity than they could in isolation.

AI Usage Statement

For this essay, I used ChatGPT 5 to help interview me and externalize the different parts of this project. I then used the same thread to help draft a detailed outline. After writing and editing, I further used ChatGPT 5 as an editor for organizational and flow support. All work in this essay reflects my thoughts and insights and was reviewed line by line.