Discouraging (Not Preventing) Unauthorized GenAI Use

Many faculty and instructors want to restrict GenAI use in some way in their courses—whether by limiting the ways AI is used, limiting AI use to particular assignments, or largely prohibiting the use of any Generative AI tools in any capacity in a particular course. Limiting AI use is challenging given powerful social forces pushing AI use and how rapidly AI technologies have integrated into our writing platforms (including Google and Microsoft). Nonetheless, there is great value in engaging in writing as a way to engage with content and develop ideas, so it’s worth taking the time to build structures that encourage students to carefully limit their AI tool use and build their own capacity as thinkers and writers.

Build Students’ Awareness and Skills

Continue to Support Skill Development

The work you already do in your classroom helps students develop traits that encourage them to engage more critically with AI tools. Continue your existing work to build students’ content knowledge (Rosbach et al., 2026), critical thinking skills (Hou et al., 2025), and intrinsic motivation to learn about the topic (Cui et al., 2026), all of which have been shown to reduce outsourcing to AI tools.

Educate Students on AI

AI literacy (Tian et al., 2025), prior knowledge of AI (Tully et al., 2025), and skepticism about AI (Beck et al., 2026) are all correlated with more cautious AI use. Helping students develop thoughtful perspectives on and approaches to AI tools can encourage them to make more informed choices. It can also be helpful to encourage students to identify factors like starting assignments last-minute that might motivate them to misuse GenAI tools and plan strategies to avoid that temptation.

Identify Examples of AI Hallucinations, Blind Spots, and Fails

While Generative AI tools are fast, they are also notoriously prone to error. While the tone and style of GenAI output can deceive novices without subject-matter expertise, experts can easily spot errors and issues in AI-generated prose. Faculty can illustrate errors in the context of your subject matter and ask students to identify errors in AI-generated prose about topics of interest to them. The internet also provides a regular catalog of spectacular and embarrassing AI failures.

Explain and Justify Your AI Policy

While it’s easy to fixate on examples where students seemed to flaunt GenAI policies intentionally, many lapses result from confusion about policies or even about what counts as GenAI. Ensuring your policy is clear and reiterating it regularly can go a long way in discouraging disallowed use. To help motivate students to follow your policy, explain how it helps students achieve your learning outcomes and connect policy decisions to the values of your field.

Build Students’ Motivation

One of the strongest predictors of academic honesty is motivation. Consider strategies that build student motivation, like allowing students to pursue topics of interest to them, providing authentic audiences, and incorporating collaborative writing or peer response. You can also address common barriers through strategies like scaffolding, low-stakes writing, and self-assessment.

Reintroduce Friction: Making Overreliance on GenAI Difficult

As noted previously, one lever tech companies use to encourage AI use is to build choice architectures that make using AI frictionless and automatic while requiring active effort to opt out. To help students choose to avoid AI tools, create assignment processes that support independent assignment completion and make AI outsourcing more difficult or less effective.

Add Specificity

Design Assignments That Require Specific Course Materials and Media 

Because AI tools scrape large swaths of writing from across the publicly available internet, they are prone to generalizations. Narrowly constructed, focused assignments that require interaction and interpretation of specific readings or materials are preferable to generic summaries or “all about” writing assignments. Inaccurate or hallucinated citations and references, or the inclusion of materials from outside the assigned content, can often reveal the presence of generative AI output.

Design Assignments Drawing on Unique Classroom Experiences

Since AI tools don’t have access to shared classroom experiences, they can’t effectively complete assignments that require students to draw on this information without substantial intellectual contributions from the student.

Add Writing Process

Incorporate Opportunities for Drafting, Revision, Feedback, and Reflection in Writing Assignments

For larger or more important writing assignments, ask students to submit early evidence of their writing processes and encourage them to take notes, create outlines, construct mind maps, and use drafts. Topic proposals, initial hypotheses, preliminary methods sections, and other scaffolded assignments help students build the competence and confidence to do their academic work. Similarly, open-ended questions in response to early drafts (particularly those that probe for greater detail, specificity, or omitted information) can engage students in writing and learning processes. 

Although GenAI tools can create facsimile drafts and reflections, they’ll often struggle to mimic authentic student writing process. GenAI reflections often lack detail and specificity.

Require Students to Document Their Research and Reading

Asking students to provide screenshots of library searches, offer UMN library permalinks to sources, use citation managers, and share annotated sources provides a durable record of engagement that is difficult to feign with GenAI tools.

Incorporate Writing Time During Class Session

Work time in class provides a secure writing environment and increases accountability around students’ engagement in the writing process.

Discouraging, Not Preventing

Of course, none of these strategies completely prevent a determined student from using AI tools. When assessing a learning outcome requires being as certain as possible that students have committed certain details to memory, (more) secure assessments like bluebooks and oral exams are certainly worth considering.

At the same time, we hope instructors will not abandon writing in their courses because of security concerns. Not everything we do in a college classroom is, or should be, aimed at summative assessment of students’ capacities: The grades aren’t ultimately as important as the opportunities we provide for learning and growth. Most people in higher education have experienced the struggle and reward of closely reading and analyzing a text, or using writing to clarify our thinking and share our expertise. Depriving students of the chance to write something valuable, meaningful, and beautiful is a cynical and counterproductive way to resist GenAI. Not every student in your courses will choose to engage deeply in reading and writing—but they all deserve that powerful opportunity.