Customizing Your GenAI Course Policy: Surfacing and Promoting Disciplinary Values

Did you know that there’s a helpful worksheet for crafting a GenAI syllabus statement (created in collaboration with Liberal Arts Technologies and Innovation Service, CEHD-Digital Education and Innovation, Academic Technology Support Services, CEI, and Writing Across the Curriculum)? Developed for joint programming on best practices for GenAI statements, the worksheet can guide instructors through customizable templates or allow instructors to craft comprehensive syllabus statements from scratch. Whether your course prohibits all uses of GenAI, permits restricted use, encourages student exploration, or requires the use of University-approved GenAI tools, these policy statements can help clarify crucial issues surrounding these tools.

But how can we help students understand the range of policies they might encounter as they travel across departments, programs, and classes? Why do some programs encourage AI applications while others steadfastly discourage their use? Does all this variation imply such decisions are arbitrary individual preferences rather than grounded and principled choices?

The advice in this blog post offers strategies for turning a run-of-the-mill policy statement into a valuable window into the academic and professional practices of your field. Taken together, this post and the worksheet can help you generate an AI policy that meets institutional expectations, demystifies some of the complexity of policy variations, and reminds students what it means to apprentice in your field.

Guidance from the Provost on GenAI Syllabus Statements

Working with a resource provided to the Senate Committee on Educational Policy, the Provost's Office recommends that all syllabi include three key pieces of information about the use of Generative Artificial Intelligence in the course, whether or not GenAI use is permitted:

1. Whether and when GenAI may be used in this course.
2. Which University-approved tools are allowable, if any.
3. Penalties for unauthorized use of GenAI and procedures for reporting

If GenAI use is authorized to any degree, then two additional statements are recommended:

4. An explanation of AI attribution and citation expectations for the course.
5. Additional information on data security, data privacy, and copyright.

Customizing the Required Materials

To make the required components of your statement more course-specific, meaningful, and motivating to students, consider these strategies:

First, connect your GenAI policy to factual, conceptual, procedural, and metacognitive dimensions of learning in your discipline. All disciplines promote ways of thinking and doing their work; make it clear where your policy aligns with these values and expectations.

For example:

• Because information literacy is an important part of the Animal Science Curriculum, students are not permitted to use AI tools for the selection or evaluation of source materials in this course.
• Because this course involves communicating with the public about animal welfare, students are permitted to use university-approved GenAI tools to identify and collect health misinformation on publicly available websites.
• Because individual reflection on your clinical experiences encourages professionalism and effective judgment, the use of GenAI for reflection is not permitted in this Nursing course.
• GenAI tools like Abridge and Nuance are now ubiquitous in clinical environments; in this course, students may consult summaries generated by these tools to recall details of their clinical experience, while being mindful of patient privacy. Note: the University of Minnesota has not authorized these tools for use in a clinical context.

Second, note the resources students will use in class and the sources of any course materials or data sets. Discussions about the materials students will read, listen to, research, discover, and create are an excellent opportunity to guide ethical use of GenAI. 

For example:

• The website data.census.gov features a built-in AI-assisted search bar that can help members of the public answer questions regarding the demography of a particular region, state, or municipality. Because all information is monitored and vetted by the Census Bureau, AI-generated search results can be considered more trustworthy than results generated by a general-purpose LLM-backed search tool or AI research assistant.
• The performance recordings you will encounter in this class are protected under U.S. copyright law and should not be uploaded to any generative AI tool. While the pieces performed may be well-worn, familiar, or even in the public domain, performances of those pieces are their own form of intellectual property.
• Course readings are available on reserve from the University Libraries. Because learning to read for information and sort relevant details is required in our major, we ask students not to use AI tools to generate summaries, even if you have no intent of submitting the generated material. Uploading course reserve materials may violate library user agreements for published resources and can have severe penalties.
• The goal of this assignment is to create a proposal as a team, not merely to divide the labor among individuals. Please obtain permission from all group members before consulting any AI tool or uploading any element of your materials. Although a GenAI writing editor might identify variations in style and tone or flag redundancies, the use of GenAI tools for editing is not required, and teammates should only use these tools with consensus.

Further Customizing with Optional Statements

In addition to the five items above, the Provost’s office notes several other potential components for developing your course policies, including statements on the value of original thinking, notes on instructors' use of GenAI, and built-in flexibility in course policies. Below are additional options to include in your course policies.

Include descriptions of core writing abilities.

Courses and programs should help students develop specific, discipline-relevant writing abilities—as articulated by departments and programs in their work with the Writing-Enriched Curriculum program. When used to augment students' capacity as writers in the field, GenAI can help them identify, assess, and demonstrate their competence in disciplinary writing expectations. Centering relevant skills and priorities helps students think critically about their use of GenAI.

Include links to professional societies, publishers, and organizations in your field that have adopted AI policies. 

Accounts from beyond the classroom emphasize the wider context of such decisions, and summaries of official policy content can help students understand the consensus of your field's experts. Purdue University Libraries have collected many academic publishers’ statements on AI.

Explicitly identify why instructors in the same program may have differing AI policies. 

Instructors are aware of how course content, the course's level and audience, and other variables and considerations in classroom practice influence AI decisions. Pointing out those differences can help students see that the policy is a decision grounded in reflective teaching practice, not an arbitrary personal preference.

For example, a lower-division Computer Science course introducing students to Python may have a very restrictive GenAI policy because learning to code is crucial to the course. After several semesters of hard work and practice, however, students should have sufficient experience to differentiate elegant, well-designed code information from functional but excessive AI code slop. Thus, an upper-division course on data structures and storage may allow for the use of an AI coding assistant for low-level tasks after students have demonstrated those skills.

Emphasize the thinking routines and practices that students will develop in the course. 

By providing detailed examples of how critical thinking or disciplinary lenses will inform students' work in class, they better understand what is lost by outsourcing work to an AI.

Include recent research on GenAI and working memory, conceptual learning, and information retention from your field. 

Students are bombarded with advertising from GenAI companies, thanks in part to the data mining of large tech companies that have identified college students as core users. Selecting research from pedagogical journals in your field can help students consider the short- and long-term consequences of outsourcing tasks to AI. While popular media and general interest higher education journals often publish studies on the consequences of AI use, information particular to your discipline or areas of expertise may be especially valuable.
⸸⸸⸸

Remember, course policies are the start of a conversation about GenAI,  not the end of an argument. Additional materials and resources are available through our Teaching with Writing Resources site, The AI Hub, and [email protected].