Course and Assignment Design

A student reaching toward a glowing AI answer box, surrounded by a circular arrow showing the vicious cycle of cognitive offloading.

The vicious cycle of detrimental cognitive offloading. Image generated with ChatGPT.

This page offers practical guidance for designing courses and assignments that discourage the misuse of generative AI (GenAI), defined as uses that undermine learning. Whether you restrict, permit, or take a mixed approach to GenAI, we hope these ideas will help you and your students meet your learning objectives.

When students treat AI as a "magic answer box," they may complete tasks without developing the knowledge and skills those tasks are intended to build. Over time, this can create a cycle of cognitive offloading: as students' foundations weaken, they become increasingly reliant on AI. Thoughtful course and assignment design can help break that cycle. For the research behind this advice, see AI and Learning.

This page covers different teaching moves. Jump to the one you need:

Motivate Your Students   |   Communicate with Your Students   |   Help Your Students Build Critical Skills   |   Use AI-Aware Assignment Design Strategies

Start with your Learning Objectives

The first step is returning to your learning objectives and examining how your assessments support them. Consider: 

  • What has changed in your discipline because of generative AI? 
  • What knowledge and skills do students need now? 
  • Which professional practices remain essential, and which are evolving? 
  • How can your assignments help students develop both subject-matter expertise and AI literacy?

Related resource

  • See strategies for writing learning objectives and designing assessments that measure student learning on the Aligned Course Design page.

Motivate Your Students

Make the purpose visible.

Motivation drives students to put in the time and effort that learning requires. The value a student assigns to a task determines how motivated they are to complete it (Ambrose et al., 2010). Making an assignment feel meaningful can be genuinely hard, but naming its purpose for students is an important start.

What knowledge and skills does an assignment build, and how will those growing capacities connect to students' personal, academic, and professional lives? Answering those questions for students, in the assignment itself, is what makes its purpose visible.

Related resources

Communicate With Your Students

Clarify AI expectations.

UO faculty are urged to have course policies on GenAI. TEP has policy examples to help. But the policy on your syllabus doesn't always match the values and experiences that students bring to your class, and in a small UO study, only 1 in 10 students describes feeling comfortable asking an instructor to clarify their expectations around AI use (see AI and Learning for more on why).

  • Share Your AI Policy as Students Begin Work. Even if you explain your AI course policy in week one, students may not remember it. Refresh their memory closer to their actual work, and mention your AI guidelines in assignment instructions. Indeed, assignments are a series of tasks—brainstorm a topic, search for sources, revise a draft. Does your stance on AI use actually shift by task? TEP's assignment template includes a dedicated section for your AI policy. See the AI-Aware Assignment Template (Word).
  • Provide the Rationale for Your Policy. Students report that understanding the rationale behind a policy is more powerful than understanding a rule to follow. A clear "why" also helps them navigate gray areas the policy doesn't cover.
  • Actively Invite Questions. Students report that asking about GenAI use can feel risky—like it signals an intent to cheat. Address that concern by creating space for them to comfortably ask clarifying questions.

Help Your Students Build Critical Skills

AI literacy and metacognition protect critical thinking.

Unstructured frequent AI use carries a risk of creating over-reliance and may be harmful to critical thinking (Gerlich, 2025; Bastani et al., 2025; Kulal, 2025). AI literacy and strong metacognitive habits protect students' critical thinking in the age of AI. Even if developing these skills is ancillary to your key objectives, simple moves like prompting students' reflection on how they're using AI make a positive difference.

Metacognitive Habits

Metacognition is thinking about thinking—including how learners plan to solve a problem, consider their progress, and draw on what they know about how they learn. Intentional pauses and reflection questions before, during, and after AI use help students attend to how they're processing new information.

Questions to pose to students (in a class activity, an assignment, or on your syllabus):

  1. Pause before starting: Do I need AI for this, or am I reaching for it out of habit, because I'm falling behind, or because I don't know where to start? Even if you decide to use a tool, this pause helps you check whether AI would genuinely support your learning here, or whether you're about to outsource the task completely.
  2. Before typing your prompt: What do I already know about this topic? This activates prior knowledge and gives you something to compare the AI output to.
  3. Evaluating the output: What will I keep, change, or reject, and why? What's still unclear, and how will I find out?
  4. Is it a product or a process? Could I explain this to someone else, when the AI output isn't around? That question helps you see whether the output is just a product, or part of a learning process that helped you achieve something the tool alone couldn't.

AI Literacy

AI literacy is "a foundational conceptual understanding of AI. It focuses on knowledge, critical thinking, and ethical awareness rather than technical skill" (Chiu et al., 2025). AI literacy isn't the same as AI competency or fluency. AI literacy is a systemic understanding of AI tools, while competency is knowing how to use a tool to complete a task. An AI-skeptic who uses no AI tools at all can still have higher AI literacy than an everyday tech-savvy user.

Related Resource

Use AI-Aware Assignment Design Strategies

Four moves for assignment design.

There's no way to make an assignment fully AI-proof. What you can do is make it AI-aware: designed around your learning objectives, with a clear view of where AI could support that learning and where it could replace it. AI-aware doesn't mean AI-allowed: an assignment that permits no AI still benefits from being designed with AI in mind.

The real goal is alignment. Whether you limit AI or invite it in, make sure your assignment design, learning objectives, and AI expectations line up, so students still put in the effort and thinking the assignment is meant to build. When that alignment is missing, AI use can tip into misuse. What alignment looks like will vary from course to course. The strategies above always matter: motivate your students, communicate transparently, clarify your expectations, and help them build critical skills. With those in place, here are four assignment design moves to try.

The Four Moves

Each move below is a way to design assignments so students' own thinking stays at the center. Start with the one that fits your course, and see how UO faculty have put it into practice. Want to see more examples? Visit the UO Faculty Examples page.

Make Purpose Visible

When students understand why an assignment exists, they're less likely to see it as busywork and outsource it wholesale to AI, and more likely to engage with it genuinely. That purpose needs to be visible at the assignment level—in other words, more specific than a course-wide AI policy.

In practice at UO

  • In JCOM 304, Damian Radcliffe's "Show Me the Money" assignment has student groups present on a media company's revenue model, with research modeled and evaluated in class and assessment weighted toward accurate, relevant citations. Purpose stays visible because the task mirrors real industry work.
  • In Computer Science, Ram Durairajan scaffolds a course from no-AI, to AI-augmented pair programming, to "force-multiply with AI," built on the premise that fundamentals still matter deeply, so students see why each stage exists before adding the tool.
Related resources

Anchor to Context

Connect the task to human experience, local context, personal interpretation, or fieldwork.

In practice at UO

  • In ENVS and Data Science, Ashley Cordes's "Neural Network" assignment asks student groups to build beaded physical representations of neural networks using art-based research methods, exploring new ways to visualize how information, memory, and connection operate within AI systems. Students can use tools like Claude for inspiration but must describe how the AI-generated visualizations compared and contrasted with their own beaded designs. They then freewrite on whether the hands-on process changed their understanding of AI systems, and what the tactile, labor-intensive act of beading reveals that stays hidden when neural networks are treated only as abstract digital technology.

Promote Process

Make the process visible: staged drafts, in-class steps, reflections, oral checkpoints. Grade the thinking, not just the artifact.

In practice at UO

  • In the College of Education, Maithreyi Gopalan pairs a course policy naming exactly where AI helps and where it's detrimental with a portfolio-based final assessment and in-class rapid rewrites aimed at what she calls "productive frustration."

Verify Learning

Create moments to see student skill directly: in-class writing, oral defenses, live problem-solving, without or beyond AI.

In practice at UO

  • In JCOM 201 and 302, Justin Francese has students choose their own event and outlets for a media-framing analysis, then document the whole research process through annotations, screenshots, a written reflection, alongside the final analysis, so the thinking is visible, not just the product.
  • In Geography, Leslie McLees built a GEO 201 rubric that rewards high-quality, evidence-based work no matter what tools were used, shifting the focus from policing AI to the criteria that actually matter.

Test Your Assignments

Run your own assignment through AI tools before assigning it. Knowing what it produces tells you where the real risk is, and where to adjust so students show the effort they need to show. Our AI Assignment Testing Guide (Word) helps with this.

Want to see this in practice? Check out how UO faculty members have put these moves to work on the Faculty Examples page.