Supporting students in the age of GenAI

Nearly every educator I know is exhausted and discouraged from dealing with ChatGPT and Generative AI (GenAI). True, some colleagues are excited about the creative potential of these new technologies, and by the role they might play in making education more accessible for disabled students and students learning in an additional language. But much more common is the complaint that students are using GenAI in dishonest ways that compromise their learning (not to mention creating a huge burden on professors, TAs and investigating deans).

Having a better understanding of why students are turning to GenAI to write their assignments can help universities more effectively reinforce academic integrity without unjustly disadvantaging the most vulnerable students.

“For students who are poor, short on time, and desperate to make the grade, the frictionlessness of GenAI can be the tipping point between academic integrity and academic misconduct.”

Studies reveal two main factors associated with student cheating. First, students are more likely to cheat when they are extrinsically rather than intrinsically motivated – for instance, when they are taking a course to satisfy a program requirement rather than out of curiosity about the course material. (If you have ever gamed your university’s online cybersecurity or hazardous materials course in order to tick the box with HR and stop receiving annoying reminder emails, you understand first-hand that extrinsic motivation can lead to corner-cutting.) Second, when students perceive cheating as common among their peers, they feel pressure to cheat too, in order to keep up. These two factors reflect, not bad eggs, but students making regrettable choices about how to allocate their available time and how to mitigate perceived risks.

These decisions are further affected by socio-economic, family or immigration status. A student who has to work 30 hours a week to stay above water or who has heavy family caring responsibilities needs to use their time very strategically. Again, this can lead to corner-cutting, especially for courses and assessments for which the student lacks intrinsic motivation. Similarly, an international student struggling academically risks losing their visa if they fail enough courses. They might take a chance on cheating as a way of mitigating that risk. The more desperate students become, the more pressure they feel to just pass the course by any means necessary.

“Like COVID-19, GenAI is with us to stay.”

We should therefore expect that non-traditional students, including low-income students, adult learners and international students disproportionately number among those accused of academic dishonesty. Further, there is evidence that racialized international students (for instance, Asian and African students, as opposed to white students from Australia or the U.K.) are over-reported for academic misconduct. When making a judgment call about an assessment’s academic integrity, some academic staff may be influenced by bias.

GenAI exploits these challenges. Buying a paper from an “essay mill” is much more expensive and much slower than getting ChatGPT to write your assignment. For students who are poor, short on time, and desperate to make the grade, the frictionlessness of GenAI can be the tipping point between academic integrity and academic misconduct.

I’m not proposing a laissez-faire approach to the rash of GenAI academic misconduct we are seeing. Universities need to defend academic integrity – both to ensure the fidelity of the academic credentials we grant and to protect students from a culture of cheating and the corresponding pressure to cheat. At the same time, we need to ensure that we are not further disadvantaging non-traditional students.

Universities’ approach to academic integrity should be holistic – not only a set of processes for reporting, investigation and punishment, but also upstream interventions that target the reasons students cheat. Here are a couple of examples: better student funding and childcare can buy students enough time to do the work themselves; well-designed curricula can reduce the box-ticking that produces extrinsic motivation. Downstream, student discipline processes should be trauma-informed, equitable and proportionate. In Canada, theft of under $5,000 does not result in deportation, but an academic misconduct finding on an assessment worth 10 per cent of a course grade can lead to a student losing their visa. As a sector, we need to rethink our student discipline systems and build in approaches that provide students a path back from misconduct.

When COVID-19 hit, the whole sector turned on a dime so that students could keep learning. We need to do the same with this latest “pandemic.” Like COVID-19, GenAI is with us to stay.