AI-Resistant Assessment Approaches
As AI use becomes increasingly commonplace, so too are concerns that students may use the technology in ways that can undermine their learning and that faculty will be able to effectively assess whether students are meeting the learning objectives for each assignment or course writ large.
What follows is a short list of ideas for how you might design or redesign assignments and assessments if you want to limit or put boundaries around AI use.
Best Practices
Transparency: Be clear about the purpose of assignments and assessments and what you expect students will learn from them.
Structures and support: AI misuse tends to happen more when students feel under time pressure or feel unable to complete the assignment. For small and large assignments alike, providing structures and support can not only encourage students not to wait until the last minute, but also affirm their belief in their capacity to successfully complete the assignment on their own, without resorting to an AI tool. For larger projects, it can be helpful to divide work into stages with check-ins and milestones along the way. Give time to explain assignments and answer questions in class and help students get started.
Clear policies: Because AI policies can differ across courses – and even within the same course – and different instructors may have different expectations for how, whether, and the extent to which students use AI, clear communication of expectations is essential. Discuss with your students what acceptable and unacceptable AI use looks like in your course and assignments. If AI is permitted, provide students with guidelines for AI attribution so they can document their use. Consider also discussing and clarifying your policies regarding potential grey areas, such as using AI to improve grammar or organization. Post these policies on Canvas and include them in your assignment instructions.
Use Cornell-supported AI tools: If you do allow students to use AI for certain tasks, encourage them to use Cornell-supported AI tools, such as the Cornell instance of Copilot, to keep information and data within Cornell’s data infrastructure.
Encourage critical approaches to AI: The AI Critical Literacy Program Canvas module is a good starting point and can be used for class discussions.
AI-Resistant Assessment Approaches
Bring writing, thinking, and problem-solving activities into the classroom
- Complete in-class activities on handouts or index cards for feedback or participation credit.
- Have students use a dedicated notebook for writing in class.
- Devote time to starting a first draft of a writing assignment in class on paper.
- Brainstorm or outline writing projects in class; consider using sticky notes, whiteboards, or online collaborative tools such as Miro.
- Use concept-mapping activities on paper, whiteboards, or with sticky notes. Students can take a photo to submit for participation credit and/or keep for their own records.
- Try end-of-class reflection activities, such as listing takeaways or key points or reflecting on how their writing assignment is developing.
- In-class quizzes, questions, or problems can be graded or used for practice.
Design local, community-based, experiential, or personalized assignments
- Consider community-engaged learning projects. For support, consult with the Einhorn Center for Community Engagement.
- Have students conduct oral interviews with experts or community members or research a local case study.
- Have students attend a local event and write a personal reflection about the experience.
- Have students collect and analyze local data.
- As a class or individually, visit a local place, such as a business, farm, or natural area, and write or record personal observations or recommendations.
- Ask students to incorporate or respond to ideas from class discussion in a written assignment.
Focus on the process
- Divide projects and writing assignments into stages with scaffolding. Turnitin now makes it easier to design multi-part scaffolded assignments.
- Bring parts of the writing or project tasks into the classroom to give students time, support, and feedback, especially in early stages of a project.
- Ask students to document their process for a project or writing assignment. This can include brainstorming notes, annotations on sources, drafts, peer feedback, and a learning reflection journal. Consider making these separate assignments that can be collected into a portfolio.
- If students work in a Google Doc shared with you, you can see the version history if needed to document their writing process. A similar approach can be used by enabling Track Changes in an MS Word document from the beginning.
- Incorporate peer review and feedback on drafts.
- Encourage students to submit revisions of assignments, along with reflections about what they have changed and what they have learned.
Use oral assessments or combine written assignments with an oral component
Examples of oral assessments include:
- Oral exams
- Required meeting(s) about student papers or projects. Checking in with groups about their projects during class time.
- Oral presentations
- Peer interviews with rubrics to practice for an oral assessment
- Video or audio recordings
- Audio reflections about the reading or a class discussion
- Audio recording of two students discussing a course topic together
- Video recording of a student explaining or teaching a topic
- Digital storytelling assignments, such as podcasts or video stories
Rebalance the weight of different assessment components
- Grade homework as complete/incomplete, provide an answer key, and reframe homework as an opportunity to practice. Give more weight to in-person quizzes, discussions, and activities.
- Grade stages of a project (early drafts, outlines, etc.) as complete/incomplete. Give more weight to a final presentation or oral discussion of the project.
- Allow students to drop their lowest scores or resubmit revisions to focus on the learning process and encourage growth over time. This can also reduce stress and time pressure.
Proctor exams and writing assessments
- Consider using Respondus: a lockdown browser for in-person proctored exams or writing activities using a laptop. Ask students to do a practice exercise before the exam day to make sure they have the program installed and working correctly.
- Consider using blue book exams or handwritten, in-class writing assessments.
Resources
This resource was informed by the following sources, and also by the Teaching Innovation Case Studies, listed below, highlighting strategies from Cornell faculty:
AI Resilient Assignments - Center for Teaching and Learning, Washington University
Redefining student assessments in AI-infused learning environments, Ncube et al. 2026
Show Your Work: Assessment in the Age of AI. Yee et al. 2026
Cornell Case Studies
- Students Turn Theory into Practice with Reading Group Guides for Middle School Youth
- Collective Discovery: Group Problem-Solving Using Student-Generated Data
- Experiencing Food Systems Firsthand with On-Campus and Local Field Trips
- Awe-Walks and Sensory Savoring Routes: Ithaca's Natural Landscapes as a Living Laboratory for a Course in Psychological Well-Being
- Oral Assessments to Reinforce and Evaluate Student Learning: A Case Study in Fluid Mechanics
- Engaging the Community: An Applied Approach to Digital Storytelling in Dietetics
- Writing about Controversies in American Politics: Using AI to Examine Counterarguments and Definitions