Build a GenAI-Resilient Course
You can’t “GenAI-proof” your class. You can create one that teaches ethical and effective use of GenAI tools.
Since generative AI made its appearance, the number‑one question I get from instructors is, “How do I GenAI‑proof my class? I’m afraid students are cheating.”
The truth: you cannot GenAI‑proof your class. But you can understand and adapt to how GenAI will impact your class; those are not mutually exclusive. There are also many good reasons to incorporate generative AI into your syllabus and your teaching methods. If we assume GenAI is here to stay, it’s better to teach students how to responsibly and ethically integrate the tool into their work and still show their own accomplishments.
Before spellcheck, being able to spell correctly was a key skill. Students looked things up in the dictionary and carefully proofread. When spellcheck became common, the bar got higher–papers were expected to be free of spelling errors. Assignments with lots of errors signaled not only poor spelling, but also that students were not using available tools. Most likely, you have read a paper with an autocorrect mistake. Now, you recognize that students still need to proofread before submitting.
Generative AI is similar–it can up the baseline for the quality of work you should be able to expect from students. However, finding ways to both accurately measure student work as well as being aware of GenAI in the classroom are not mutually exclusive. You can do both, creating a classroom resilient to GenAI’s impact.
Consider How GenAI is Used in Your Discipline
First, think about how GenAI fits your discipline. Is it already part of professional practice? Will it be? This may impact how you decide to integrate GenAI into your classroom, or not. If you anticipate that GenAI will be an important part of students’ future careers, consider ways to integrate it.
Try Your Own Assignments in GenAI
Don’t assume that your assignments aren’t GenAI proof. Try it out. Paste your assignments into a GenAI model and see how it responds.
- Note what it does well (summaries, comparisons) and where it struggles (nuanced critiques, original synthesis, local context, personal experiences).
- Keep samples as “GenAI baselines” you can show students and design against.
Do this for several of your writing assignments and your main assessment. You’ll quickly see the likely misuse points. You also may decide that your assignments are relatively GenAI proof already.
Align What You Assess to What Only Humans Can Show
Think carefully about what you are assessing and compare that to what GenAI is good for. GenAI is great for retrieving facts, summarizing, and comparisons. It’s not good at critical thinking, creative solutions, and next-step analysis. The most vulnerable assignments to GenAI are those where critical thinking is the least important. Multiple choice exams are frequently at risk of GenAI because more often it’s recall being assessed, rather than how to apply information.
| GenAI excels at: | Humans must still show: | Design moves: |
| Rapid drafting, summarizing, formatting | Reasoned judgment, orginality, voice | Require rationale memos (“why these choices?”), reflections, and source-to-claim chains |
| Surface-level comparison, definition recall | Transfer to new contexts, evaluation, ethical reasoning | Use novel cases, local data, and constraints that require interpretation |
| Polishing prose and style | Domain knowledge, method selection, data work | Grade the process: data collection, analysis steps, method justification |
| Idea generation (lists, outlines) | Curation, prioritization, feasibility | Have students who selection criteria and trade-off analysis |
Build GenAI-Resilient (and Inclusive) Assessments
The best way to beat GenAI? Join GenAI. Invite appropriate use of GenAI while requiring artifacts that demonstrate learning. Encourage students to iterate on what GenAI can do.
- GenAI‑baseline + revision: Students generate an outline or first draft with GenAI, submit the prompt and output, then revise extensively. They submit a revision memo explaining what they kept, changed, and why.
- Local/unique data: Collect real data (observations, interviews, lab measurements) and analyze it. GenAI can help organize, but the data work is theirs.
- Student‑built study tools: Have students prompt GenAI to create low‑stakes practice quizzes, flashcards, or explainers—then critique and correct them.
- Feedback loops: Let GenAI point out “3 places to improve,” require students to test the advice, and reflect on what actually helped.
If GenAI isn’t appropriate for your type of assessments, consider the following:
Controlled recall when needed: For essential memorization, use in‑person closed‑resource quizzes or oral checks. Keep them short and purposeful.
Alternative media: Since GenAI is strong at generic writing, shift some outcomes to video explainers, diagrams, demos, or portfolios. Find different ways for students to apply their knowledge.
What Not To Do
Don’t avoid it. By ignoring GenAI, you will likely become frustrated and end up dealing with the consequences of GenAI use that you didn’t approve. Designing “gotcha” exercises to catch students using GenAI is just going to make students try harder. Even GenAI-detectors are fallible as they can mislabel writing.
If you are unsure about GenAI, if you don’t have the time or mental capacity to deal with GenAI, you can still find ways to create space for it.
Set Clear Guidelines in Your Syllabus
The university has several resources for developing a clear set of guidelines for the use of GenAI in your classroom.
