Dr. Danielle Condry is an associate professor and graduate director in the Department of Plant Pathology, Microbiology, and Biotechnology at North Dakota State University.
A few months ago, I asked a room full of faculty a simple question: If AI can complete the assignment, what exactly are students learning by completing it? The room got quiet. Not because the answer was obvious, but because many of us are wrestling with the same tension. We know AI is here to stay. We know our students are using it. We know many of the careers they are preparing for will require them to work alongside AI tools. At the same time, we know something else: learning requires effort. If AI can instantly generate answers, essays, and solutions, where does that leave the struggle that is so essential to learning?
The more I work with AI, the more convinced I become that the answer lies not in the technology itself but in how we design learning experiences around it. If the goal is simply to produce a final product, AI will often outperform students in speed and efficiency. But if the goal is thinking, decision-making, creativity, reflection, and revision, AI becomes something very different: a thinking partner.
As educators, we know people learn best when they operate within their Zone of Proximal Development (ZPD)—the space between what they can do independently and what they can accomplish with support. Learning rarely happens when tasks are too easy, but it also stalls when challenges become so overwhelming that students shut down. The sweet spot lies in productive struggle: that uncomfortable place where students wrestle with ideas, make mistakes, revise their thinking, and gradually build expertise.
Unfortunately, many students interpret struggle as evidence that they are failing. When they say, “I don’t know what to write,” “I don’t know where to start,” “This is taking too long,” or “I don’t get this,” they often view those moments as barriers to learning. Those moments are signs that learning is happening. The challenge for educators is helping students navigate that discomfort without removing the cognitive work altogether. Used thoughtfully, AI can help students move through productive struggle while preserving the thinking that leads to growth.
AI also presents students—and faculty—with an important question: What am I avoiding when I turn to AI? Sometimes the answer is perfectly reasonable. We may need a starting point, a second opinion, or help organizing our thoughts. Other times, however, we may be trying to bypass the uncertainty and effort that often accompany learning. The struggle itself is not the goal, but neither should we eliminate every challenge we encounter. In higher education, we often treat efficiency as inherently desirable, yet developing expertise, creativity, judgment, and comfort with ambiguity is rarely efficient. The question is not whether AI can help students work faster. It can. The question is whether working faster is always the same thing as learning more deeply. I am not convinced that it is. Sometimes the moments that feel slow, frustrating, or uncomfortable are exactly where growth occurs.
Consider a student assigned to write a paper. One option is to enter the assignment prompt into an AI tool, copy the response, and submit it. Could they do that? Certainly. Would they learn much? No. Now imagine a different approach. The student uses AI to brainstorm possible angles for the paper. They ask it to identify assumptions they may be making. They request examples of counterarguments. They use it to help organize their ideas, identify gaps in their reasoning, and suggest questions they may not have considered. Then they write the paper themselves, revise it, and perhaps even use AI to receive feedback on clarity, organization, or tone before making final improvements.
At this point, some readers may be thinking, Couldn’t a classmate, tutor, or instructor provide that same support? Absolutely. In many cases, those human interactions are preferable. However, they are not always available when a student is working at 10 p.m. the night before a deadline, balancing a job, family responsibilities, or a full course load. Faculty office hours are limited. Peer support varies. In large-enrollment courses, individualized guidance can be difficult to provide at scale. AI can help fill some of those gaps by offering immediate feedback, prompting reflection, and helping students get unstuck. The key distinction is that AI should support the thinking process, not replace it. What is important here is that the student is still doing the intellectual heavy-lifting. AI may be helping them brainstorm, question assumptions, organize ideas, or get unstuck, but it is not doing the learning for them. The thinking, learning, and ownership still belong to the student. Most importantly, the voice remains their own.
That last point is one I find myself returning to often. One of my biggest concerns about student AI use is not cheating. Rather, it is conformity. Generative AI is exceptionally good at producing average responses. It generates language that is coherent, polished, and statistically likely. In many ways, it produces exactly what we would expect. But creativity rarely emerges from what is expected. Original ideas often come from taking risks, making unusual connections, challenging assumptions, and bringing unique experiences into the conversation. When students allow AI to do all of the thinking, they risk outsourcing the very things that make them valuable. They should not want to outsource the parts of learning that make them uniquely human.
Their creativity.
Their judgment.
Their curiosity.
Their perspective.
Their voice.
These qualities allow students to contribute something new to the world. They are also the qualities employers consistently seek and that distinguish meaningful work from average work. AI should amplify those qualities, not flatten them into the average. I often tell students that the goal is not to sound like ChatGPT; the goal is to sound like themselves. AI can help them, but it should not determine their final answer. Students deserve the opportunity to discover what they think before turning to a machine to generate an answer for them. This is why assignment design matters so much.
Rather than focusing exclusively on the final product, I believe we should increasingly focus on the process students use to arrive there. One effective strategy is scaffolding assignments into smaller stages that make thinking visible. Instead of assigning a final paper, we might ask students to first develop a topic, generate questions, evaluate sources, create an outline, submit a draft, reflect on feedback, and explain the revisions they made. Each step provides opportunities for students to engage in meaningful thinking while documenting and justifying their decision-making process.
The same principle applies to critical-thinking assignments. Rather than asking students to provide a single answer, we can ask them to explain their reasoning, compare alternative approaches, defend their decisions, evaluate strengths and limitations, and reflect on how their thinking evolved over time. When assignments emphasize reasoning rather than just answers, students are far more likely to engage in the cognitive work that promotes learning.
Of course, scaffolding creates its own challenges. Breaking assignments into multiple stages often means more pieces for students to complete and more pieces for faculty to review. Asking students to document their reasoning, reflect on their AI use, or explain their decision-making can quickly become burdensome if we are not careful; there needs to be a balance, and I want to acknowledge that some of that will be context-dependent (e.g., large enrollment vs. small enrollment). If the goal is to make thinking visible, we make that choice when it serves a clear educational purpose.
That does not necessarily mean every stage requires extensive instructor feedback. Rubrics can help students self-assess their work before submission. Structured peer review can provide valuable feedback while helping students learn to evaluate quality work. Some checkpoints can be graded for completion rather than accuracy, while others may warrant more detailed feedback. The key is to strategically identify moments where feedback, reflection, and revision are most likely to support learning. As with any instructional strategy, thoughtful implementation matters more than the strategy itself.
Transparency also matters. Students are more likely to engage meaningfully when they understand why we are asking them to do something. When I talk with students about AI, I try to frame the conversation around learning rather than compliance. My goal is not to prevent them from using AI. My goal is to help them learn how to use it effectively. I want them to develop judgment. I want them to recognize when AI is helpful and when it is misleading. I want them to learn how to critique AI-generated information rather than simply accepting it. Most importantly, I want them to leave my course with skills that will remain valuable even as the technology continues to evolve.
Our graduates will enter workplaces where AI is commonplace. They will not be rewarded for pretending it does not exist; they will be rewarded for using it thoughtfully, responsibly, and strategically. Our responsibility is to prepare them for that reality while continuing to cultivate the human skills technology cannot replace. The goal is not less thinking. The goal is better thinking.
MEET THE AUTHOR
Dr. Danielle Condry is an associate professor and graduate director in the Department of Plant Pathology, Microbiology, and Biotechnology at North Dakota State University. She has a passion for curriculum development and has led efforts to align microbiology and biotechnology programs with national standards and workforce needs. Her research interests include discipline-based education research, with projects on assessing student learning through concept inventories, implementing community-engaged learning in science curricula, and exploring equitable grading strategies in large-enrollment courses. Danielle is also committed to student success, mentoring graduate students, and fostering equity in education. Beyond academia, she enjoys outdoor activities, yoga, reading, gardening, cooking, traveling, and family life with her husband and daughters.