(8) AI cheating in education: What can we do right now? | LinkedIn
Recent reports, including a notable article in The Guardian, by Caitlin Cassidy have shed light on a growing concern: students are seemingly using AI tools inappropriately to complete their academic work in increasing numbers. While this issue is particularly prominent in higher education, it's also
Know your students: High-quality learning is fundamentally relational, not transactional, despite higher education looking increasingly transactional in nature (I won't rant about that issue here). While getting to know students can be challenging in large cohorts and/or with a high reliance on sessional staff, finding ways to connect with students individually can make a significant difference in promoting academic integrity and provide insight into the individual trajectories students are on in their learning. Yes, I know, much easier said than done.Be transparent about AI use: Whether using a system of "lanes," like Professor Danny Liu’s two-lane approach, the multi-lane highway approach outlined by UNSW’s Professor Alex Steel, the AI Assessment Scale developed by Leon Furze, Dr Mike Perkins, Dr Jasper Roe SFHEA and Dr. Jason MacVaugh or another framework, be absolutely explicit about what constitutes appropriate and inappropriate use of AI in your units/subjects. My sense is that coordinators are best placed to make these calls, which takes us back to the awareness-raising piece above. Clear guidelines can help students navigate this new terrain ethically but this kind of guidance is often lacking (partly because we have been trying to figure all this out, of course).Ask students to show their working: If students are permitted to use AI tools, require them to document their process. The calculator analogy doesn’t work with generative AI for a range of reasons but ‘show us your working’ is a useful heuristic here, as it was when calculators appeared on the scene. This approach could include sharing the prompts they used, the outputs they received, and how they incorporated this information into their final work. This approach not only discourages misuse but also helps students develop critical skills in working with AI and gives us some insight into how these tools can be used in the tasks we assign.Engage in conversations with students as assessment: Consider incorporating more oral assessments or discussions into your assessment. While this may be challenging in large cohorts, even small-scale implementation can provide valuable insights into students' understanding and thought processes that may not be evident in written work alone. I have resisted this one because of the challenges I face in implementing this approach in a cohort of 250 students, but I have changed my mind on this and will give it a go. I have been convinced by the argument that we can learn a lot more about how a student is going in a 15-minute chat with them than in spending an hour or more looking at the distant echoes of their progress in a written artefact.