GenAI and Academic Integrity in Online Courses
Nick Gaspar is director of online and digital education at the University of Michigan-Flint. Tazin Daniels is an associate director at the Center for Research on Learning and Teaching.
If you spend any time talking with online instructors right now, you’ll hear a version of the same question:
How do I prevent students in my online course from cheating with generative AI?
It’s an understandable concern. These tools are fast, easy to access, and often produce work that looks polished on the surface. In an online course, where you don’t see the process unfold, that uncertainty can feel amplified. You’re left wondering who actually did the thinking behind what was submitted.
That anxiety is real, and it comes from instructors who care about their students and the integrity of their courses. When we focus on stopping AI, we center the tool and shift our energy toward control. A more useful question brings the focus back to the learning environment:
What conditions are we creating that make cheating in this online class feel like a reasonable option?
This doesn’t excuse misconduct. It does push us to examine course design, workload, clarity, connection, and stakes. If a student believes no one will notice their effort, the assignment feels disconnected, or the fastest path to survival is outsourcing the work, generative AI becomes a convenient solution.
Proposal
A durable response to generative AI starts with understanding why students cheat in the first place.
Cheating often functions as a signal. It can point to pressure, confusion, disengagement, or a lack of perceived relevance. Students enter our courses with different constraints and motivations. A student balancing work and caregiving may use AI to keep up. A student who feels lost may use it to fill in gaps. Another may see it as a shortcut when the task feels transactional.
These differences matter.
This is why the conversation needs to center on instructional and relational design.
Instructional design shapes how learning unfolds. Are expectations clear? Do assignments build over time? Are there opportunities for feedback before high-stakes work? Relational design shapes whether students feel visible. Do they believe their effort matters? Do they feel connected to the course and the people in it?
When clarity and connection are present, students are more likely to invest in their own work. The focus shifts toward building conditions that support integrity.
Motivation and Design Responses
Disconnection
Many online learners experience courses as a series of isolated tasks. They log in, submit work, and move on. Over time, the course can feel transactional, and in that environment, integrity becomes abstract. If no one feels present in the process, using AI can feel no different than using any other tool.
Design Tip 1: Build connection early and intentionally
You can shift away from this dynamic without overhauling your course. A few purposeful design choices go a long way.
- Use structured introductions that connect to the course. Not a generic ‘post your bio’ prompt, but something that invites students to share their experiences, perspectives, or questions tied to the course topic itself.
- Be visibly present in small, consistent ways. This could include short weekly videos, targeted feedback, or structured check-ins with students.
- Design interaction that requires students to engage with each other’s thinking. Think of this in terms of responsibility to a group rather than themselves. When students engage with each other they see their work as part of a shared process.
- Create low-stakes opportunities early so students can participate without pressure. These interactions create a foundation that can carry forward into more complex work later.
None of these strategies eliminate misconduct on their own. They do change the experience of the course in ways that make disengagement less appealing and investment more natural.
Confusion About Boundaries
Students are often trying to figure out where the line is. Using AI to revise writing may feel similar to using spellcheck. Generating ideas or outlines can feel like part of the process, but when expectations aren’t clear, students rely on their own judgment.
Design Tip 2: Make expectations explicit and a part of your course
- Write AI guidelines in plain, direct language that students can actually interpret. Avoid vague phrases like “use responsibly.” Instead, say something like: “You may use AI to help revise grammar and clarity, but the ideas, structure, and arguments must be your own.”
- Go beyond listing rules. Walk through examples that show what acceptable and unacceptable use looks like in your course. For example: “Asking AI to suggest alternative wording for a paragraph you wrote is acceptable. Asking it to generate a full response to a discussion prompt and submitting it is not.” Tie these examples directly to your assignments.
- Invite questions and normalize uncertainty so students feel comfortable asking where the line is. Build this into the course. A short discussion prompt or anonymous question form early in the term gives students space to ask, “Would this be okay?” before they’re in a high-pressure situation.
- Address gray areas early, before students encounter them on high-stakes assignments. Spend a few minutes during the first major assignment walking through common scenarios. Show them what borderline use looks like and how to make better decisions before it counts for a grade.
Perceived Busy Work
When assignments feel disconnected from meaningful learning, students look for the fastest way to complete them. If a task feels generic or easily handled by AI, engagement can drop.
Design Tip 3: Make assignments purposeful and connected
- Be transparent about why the work matters. Take a moment to explain what the assignment is designed to help them practice and develop. Students are more likely to engage with it if they understand the purpose. The TILT framework is a great way to approach this.
- Develop more authentic assessments that connect to real contexts. This could include applying concepts to a current issue, a professional setting, or a scenario they might realistically encounter.
- When students can see themselves in the work, it becomes harder to treat it as disposable. Build in choice where possible. Let students connect topics to their own field, interests, or goals.
- Incorporate elements that require personal or local application, along with opportunities to show progress over time. Drafts, revisions, or iterative steps make the work feel more grounded in their own thinking.
It’s also worth asking a simple question as you design your course: would I want to do this assignment? If the answer is no, it’s likely that students will feel the same way.
Overload and Pressure
Many online students balance competing responsibilities. When time is limited, AI can feel like a way to stay afloat. Under pressure, AI output can reinforce perfectionism.
Design Tip 4: Build structure that supports progress over perfection
- Scaffold assignments so students aren’t carrying the full weight of a task all at once. Chunk larger projects into smaller pieces with clear checkpoints along the way.
- Use milestones to pace the work. This helps students stay engaged over time and reduces the temptation to complete everything at the last minute. A simple example is opening your course modules one at a time on scheduled dates.
- Build in revision cycles. Give students opportunities to improve their work based on feedback rather than expecting a single, high-stakes submission.
- Normalize drafting as part of the process. When students see that strong work develops over time, it shifts expectations away from immediate perfection. For example, instead of collecting a single final paper, have students submit a rough draft or outline first, receive brief feedback, and then revise before the final submission.
An Honest Look
Before we wrap up, it’s worth taking an honest look at why generative AI is so appealing to students in the first place.
It’s fast and available whenever they need it, including at 2 a.m. when a deadline is quickly closing in. It doesn’t judge or get frustrated when students need a concept explained again. It produces something that looks polished, even when the student isn’t fully confident in their own understanding. That combination is hard to compete with, especially in online environments where students are often working alone.
At the same time, there are real tradeoffs that students don’t always recognize in the moment.
Overreliance can lead to skill atrophy. Students miss opportunities to practice and develop their own thinking. Over time, they stop building the capabilities the course is designed to support. It can also create false confidence. A polished response can give students the impression that they understand the material more deeply than they actually do.
And perhaps most importantly, it removes the cognitive struggle where learning happens. Working through confusion, making mistakes, and refining ideas is part of how understanding develops. When that process gets outsourced, the learning often goes with it.
These dynamics matter for course design. You can’t design effectively around generative AI without first acknowledging why students are drawn to it.
Takeaways
This work ultimately comes down to the environment you create.
Academic integrity in online courses extends beyond policy statements. It shows up in course structure, communication, and student experience. Students respond to clarity, purpose, and connection.
You won’t eliminate the temptation to cheat for every student. What is within your control is reducing the conditions that make cheating feel necessary or reasonable.
A fully “cheat-proof” course isn’t realistic. There will always be ways for students to work around a system if that is their goal. Chasing that outcome can pull attention away from the parts of teaching that have the greatest impact.
What is within reach is a course that students want to engage in. One where expectations are clear, their presence matters, and the work feels meaningful.
Ask yourself the following:
If you were a student in your own course, would you feel invested in the work or focused on just getting through it?
