Crafting Assessments That Prevent AI Cheating

(Originally titled “Designing Plagiarism-Resistant Assessments”)

In this Educational Leadership article, Torrey Trust and Robert Maloy (University of Massachusetts/Amherst) empathize with teachers trying to decide whether to confront students who have submitted work that looks like it had an AI assist. Rather than playing cat and mouse with academic cheating, Trust and Maloy suggest using the TRUST model to create assignments and assessments that reduce the likelihood of students taking AI shortcuts.

• Transparency in rationale, task, and criteria – “When students understand the purpose behind an assignment,” say the authors, “they’re more likely to value the process and less likely to rely on AI tools to shortcut their learning.” The Transparency in Learning and Teaching website www.tilthighered.com/resources has a number of exemplars. Teachers should also clarify the rules for use and misuse of GenAI tools.

• Real-world application of knowledge – Students have lots of experience with “disposable” assessments, say Trust and Maloy – a paper that’s written, graded, and thrown away. But if a project has relevance to their lives and the world around them, students are more likely to engage intellectually and keep it after it’s been graded. In Maloy’s methods class for history teachers, students contribute to a Wikipedia page https://bit.ly/resourcesforhistoryteachers.

• Universal design for learning – Students often blame themselves when they do poorly on an assessment – when it’s actually the design of the assessment that was at fault. UDL encourages teachers to assign work that allows all students to showcase their knowledge and skills. The CAST website has suggestions: https://udlguidelines.cast.org.

• Social knowledge construction – All learning has the potential to be social, say Trust and Maloy, “whether it involves reading, listening, watching, talking, writing, or reflecting on what others have said and done.” The key is designing learning experiences that incorporate a group dynamic – for example, groups of students working on a blog post, newsletter, or podcast series and sharing it with the local community.

• Trial and error – “Learning is a process of making mistakes and using them to improve,” say Trust and Maloy, but students too often conclude they are “bad” at something and avoid the productive struggle needed to improve. Teachers can help shift this mindset by assigning a series of low-stakes tasks, normalizing errors, giving lots of encouraging feedback, and reframing failures as steps toward mastery.

The Marshall Memo Admin - Issues