In the few breathless months since launching, ChatGPT has been called “a plague on higher education” and a “superspreader event” no less dangerous than COVID, albeit “one that threatens our minds more than our bodies.” Institutional reactions have ranged from outright school bans, to more open minded responses, recognizing a future in which many professions will benefit from the assistance of AI ‘co-pilots.’
The reality is, it’s unrealistic and arguably too simplistic to cordon off something as powerful as AI. Instead, as with any disruptive innovation, the rise of ChatGPT should challenge us to examine our current approaches to teaching and learning, including the issue of student cheating.
In praise of cheating
Whenever the fear of cheating arises—whether plagiarism or the wholesale outsourcing of assignments—we should first seek to understand. Why do students feel compelled to cheat? Are ChatGPT detectors and internet disabled computer labs really the only solution?
In the relative infancy of my career, I came across an early version of Russ Hunt’s, “In Praise of Plagiarism.” Among other things, Hunt argues that the more students view assessments and assignments as hoops to jump through, the more likely they are to fall for the allure of essay mills and other shortcuts. This is especially true when the primary motivation for students is securing the grades to progress in their chosen major.
Cheating is ultimately symptomatic of a design flaw: the challenge for educators is to cultivate genuine interest in the topics we teach and the knowledge and skill we seek to nurture so that students are motivated to do the work. If I want to learn to play guitar, cheating would be the last thing on my mind. It’s counterproductive because my motivation is intrinsic.
Of course, instructors often teach courses students are required to take. Steven Mintz, a professor of history at the University of Texas at Austin, accepts that many of his students view his history course as a “box-checking exercise.” Nevertheless, he feels a sense of duty to try and alter this viewpoint by “demonstrating that history isn’t just one damn fact after another.” Connecting what we teach to student interests, their lived experience and the world outside of higher education is foundational to motivation and engagement.
Assessments for learning
Our reliance on high-stakes assessments is hardly helping. Reducing the criticality of any one test through frequent, low-stakes assessments can lessen the incentive to cheat, particularly when assessments are designed to provide value to the student along with meaningful feedback. Knowledge, as Hunt suggests, needs to be “continuously reformatted, reconstituted, restructured, and exchanged in new forms.” Our assessment practices should offer the variety and flexibility to allow students not just to apply knowledge and learn from mistakes, but to reflect and connect learning to their own goals and aspirations.
Happily, there are many benefits to this approach. Not only are frequent assessments supported by our best understanding of how people learn, they contribute to increased overall engagement, foster metacognition, and improve learning outcomes for all students. Formative and low stakes assessments can also deliver the real-time data instructors need to support learners. By pinpointing what’s working and what isn’t, educators are better equipped to modify where and how we spend our time with students, allowing us to make iterative improvements in course design.
ChatGPT is just the latest in an accelerating pace of change driven by technology that will alter how we live and work in unpredictable ways. We will have to come to terms with the challenge of regularly and rapidly modifying courses and curricula, that much is true. But while cheating may take on new forms, it is nothing new. The good news is we can design away to a very high degree the impulse to cheat. And the best evidence from learning science tells us how.