The question isn't whether students can use AI on your assignments. They already can. The real question is: have you designed your assignments with that reality in mind?
That's what the ARAD framework is for. AI-Responsive Assignment Design gives you a structured way to think about every assignment you create—not by banning or ignoring AI, but by making intentional choices about how it fits into the learning experience.
The Core Idea: Intention Over Reaction
Most faculty fall into one of two camps: they either ignore AI and hope for the best, or they try to police it after the fact. Neither approach serves students well.
ARAD asks a different question: Before you assign anything, decide what role AI should play. Every assignment falls somewhere on a spectrum:
- AI-Resistant: Designed so that AI cannot meaningfully complete the task. Useful when you're assessing foundational knowledge, personal reflection, or skills that require human presence.
- AI-Permitted: AI can be used as a tool, but the assignment is designed so that the learning still happens. Students must add significant value beyond what AI produces.
- AI-Integrated: AI use is required or expected, and the assignment specifically assesses students' ability to work with AI effectively and critically.
AI-Resistant Design: When and How
Some learning objectives genuinely require AI-free work. You want to know that students can perform fundamental calculations, articulate their own analysis, or demonstrate personal reflection. For these situations, ARAD offers practical strategies:
- Handwritten or in-class work. The simplest approach. Timed, proctored writing eliminates AI assistance entirely.
- Oral assessments. Ask students to present and defend their work. AI can write a paper, but it can't stand in front of you and answer follow-up questions.
- Hyperlocal prompts. Design assignments around class-specific discussions, local events, or personal experiences that AI has no context for.
- Process portfolios. Require multiple drafts with documented revision decisions. This makes the process the product.
AI-Integrated Design: Making AI the Assignment
Some of the most powerful learning happens when students use AI and then critically evaluate what it produces. Consider these approaches:
- AI as first draft. Students generate an initial response with AI, then must identify errors, fill gaps, add citations, and revise. The learning is in the critical evaluation.
- Prompt iteration. Give students a problem and ask them to document how they refined their prompts to get better AI output. This teaches both the subject matter and prompt engineering.
- AI debate. Have students argue against an AI-generated position. This requires deep understanding of the material to find weaknesses in an articulate but potentially flawed argument.
- Comparative analysis. Students generate AI responses from multiple tools (ChatGPT, Gemini, Claude) and analyze the differences. This builds critical evaluation skills and subject-matter expertise simultaneously.
Examples Across Disciplines
English Composition: Instead of "Write a 5-page analysis of Beloved," try: "Use AI to generate an analysis of Beloved. Then write a 3-page critique of the AI's analysis, identifying what it got right, what it missed, and what it fundamentally cannot understand about the text."
Biology: Instead of "Describe the process of cellular respiration," try: "You're a teaching assistant and a student asks you to explain cellular respiration. Record a 3-minute video explanation using only a whiteboard. No notes, no AI, no script."
Business: Instead of a standard case analysis, try: "Use AI to generate a strategic recommendation for this case. Then conduct a risk analysis of the AI's recommendation, identifying at least three assumptions the AI made that may not hold in practice."
Getting Started with ARAD
You don't need to redesign every assignment at once. Start with one:
- Identify the learning objective. What should students be able to do after completing this assignment?
- Test it against AI. Run your current assignment through ChatGPT or Gemini. How good is the output? If AI can produce an A-quality response, your assignment needs redesigning.
- Choose your position on the spectrum. Should this assignment resist, permit, or integrate AI?
- Redesign accordingly. Use the strategies above to align the assignment with your chosen position.
- Communicate your expectations. Tell students explicitly what AI use is allowed for this specific assignment.
The free ARAD training walks you through this entire process with hands-on exercises. It takes about 45 minutes and you'll leave with a redesigned assignment ready for your next class.