In the spirit of “what I did this summer,” I wanted to share a quick reflection from CTLE’s summer book club discussion of The Opposite of Cheating: Teaching for Integrity in the Age of AI. The book helped us look beyond the familiar cycle of detection, suspicion, and punishment and think more deeply about what academic integrity is really meant to protect: learning. One key takeaway was that the opposite of cheating is not simply compliance; it is learning, and the conditions we create to support that learning (Gallant & Rettinger, 2025).
Students are still responsible for their choices. Academic honesty matters. Ethical use of tools matters. Original thinking matters. But if our only response to AI is to catch misconduct after it happens, we miss a chance to look upstream. Long before students click submit, our curriculum and assessment choices have already shaped how they understand the purpose, value, and expectations of the work.
When Students Reach for AI
Students don’t usually set out to violate academic integrity policies. More often, they are overwhelmed, confused, disconnected from the purpose of the assignment, afraid of failure, short on time, or convinced the work is just another box to check. None of that excuses cheating, but it does remind us that course design is part of the integrity conversation.
Generative AI did not create the academic integrity problem. Gallant and Rettinger suggest it has only exposed design issues that were already there. Generic assignments are especially vulnerable to generic shortcuts. If a prompt could be answered by anyone, in any class, with little connection to course materials or context, AI can probably generate a passable response in seconds. That does not mean we need to throw out every assignment and start over. Most of us do not have the time, energy, or caffeine supply for that. It does mean we can look at our assessments with fresh eyes.
For example, asking students to “write a paper about the branches of the criminal justice system” may produce relevant work, but it also invites broad, surface-level responses. A stronger version might ask students to apply course concepts to a specific case, trace how the case moves through law enforcement, courts, and corrections, explain the responsibilities of each branch, and evaluate where discretion or decision-making could affect the outcome. That kind of assignment is not AI-proof, very little is, but it is harder to complete well without understanding the course concepts.
Show the Work
The goal is not to create assignments technology cannot touch. A better goal is to design work where students’ judgment, reasoning, context, and their process matters. That might mean asking students to submit a planning note, complete a draft checkpoint, develop an annotated source list, explain their research choices, or connect their final answer to a class activity, case study, lab, discussion, or reading. These design choices do more than discourage shortcuts. They also help us see student learning more clearly. When students explain their thinking, we get better evidence of what they understand, where they are struggling, and where we may need to intervene.
Lower the Cliff, not the Rigor
High-stakes, one-and-done assignments can create pressure, especially when students are already behind or unsure of expectations. We can reduce that pressure without lowering rigor by building in smaller steps before major assessments.
Practice activities, draft checkpoints, peer review, instructor feedback, and self-assessment all help students see the assignment as a process rather than a cliff. They also give us earlier opportunities to intervene before the final submission, not after.
Talk About AI Directly
Generative AI is not going away, and avoiding the subject will not make it disappear. Students know these tools exist, and many are already using them in some form.
If we want students to make ethical decisions, we need to be clear about expectations. That means explaining why some tasks must be completed independently, when AI use is prohibited, and when limited use may be acceptable. It also helps to acknowledge that policies may differ by course or assignment because the purpose of the learning task differs.
The key is helping students understand the why.
Make the “Why” Visible
Sometimes students disengage because they do not understand why the work matters. Faculty may see the connection clearly, but students may not. A brief “why this matters” statement can help students see the assignment as practice rather than busywork. For example:
This assignment is not just asking you to summarize a theory. It is asking you to practice using theory as a decision-making tool. That skill matters because professionals are often expected to explain not only what they decided, but why they decided it.
Transparency does not need to be lengthy. Even a few sentences can clarify the purpose of the task, the skill being developed, and how the work connects to course outcomes, future coursework, employment, civic life, or professional judgment. Transparent assignment design has also been linked to gains in student confidence, retention, metacognitive awareness, and workplace readiness (as cited by TILT Higher Ed, para. 2).
A Curriculum and Assessment Issue
This is where the conversation connects curriculum to assessment. If assessment is supposed to tell us what students learned, then our assignments need to make learning visible. In the AI era, that may mean fewer assignments that ask students to simply produce polished answers and more assignments that ask them to show the thinking behind the work. It also means revisiting alignment. Are our assessments measuring the outcomes we care about? Are students practicing those skills before they are evaluated? Do our rubrics reward thinking and application, or only the final product? Are we assessing what matters most, or what is easiest to grade? These are not new questions, but AI is making them harder to ignore.
Our PAIR assessment framework gives us a practical way to respond. We Prepare by designing meaningful work and clarifying expectations. We Assess what students can actually do. We Intervene when students struggle, disconnect, or misunderstand the task. Then we Reassess to evaluate whether our changes helped students engage more deeply and honestly.
Detection tools and policy statements may have a place, but they cannot carry the whole integrity load. The more promising work may be quieter and more familiar: clearer outcomes, better-aligned assessments, meaningful feedback, transparent expectations, and assignments that help students see the value of doing the work themselves.
References
Gallant, T. B., & Rettinger, D. A. (2025). The opposite of cheating: Teaching for integrity in the age of AI. University of Oklahoma Press, Norman Publishing Division of University.
TILT Higher Ed. (n.d.). Transparency in Learning and Teaching. https://www.tilthighered.com/about/about-tilt
