Questions? Please contact Cara Metz or Yvonne Lozano.
CLASSROOM MANAGEMENT
GENERATIVE AI AND ACADEMIC INTEGRITY: NAVIGATING THE GREY AREAS Alaina Pascarella, Manager, Academic Quality Services
Over the centuries, technological innovations have impacted how we view and engage in the process of teaching and learning and can sometimes leave those in academia feeling as though we are on unsteady ground. Generative artificial intelligence (AI) is one such technological advancement where there is uncertainty about its potential to impact, both positively and negatively, how we approach academic writing as a learner and a teacher. Naturally, one of the major concerns
of generative AI usage is how to maintain academic integrity in the classroom when the tool can seemingly compose written work to address classroom prompts. The uncertainty of the impact on academic integrity in the classroom is compounded by a lack of definitive evidence of generative AI usage and calls for us to consider a nuanced approach to generative AI and academic integrity in order to navigate the shades of grey of this new normal: we must balance academic integrity concerns with the need to train students to utilize tools which may impact their future work. What Do We Mean by Generative AI In order to better understand why there are concerns regarding academic integrity and generative AI, we must first briefly explore the evolution of artificial intelligence and generative AI. Previously, when we discussed the concept of artificial intelligence, we were talking about machine learning that could predict patterns in data and make predictions based on that data (Zewe, 2023). We have used this type of artificial intelligence for some time in our Signalz predictors. Demographic and student engagement data was input into a system, which would make predictions about the potential need for intervention by the faculty to help the student be successful in the classroom. Generative AI expands on this machine learning capability, utilizing that same data set to create new data rather than make predictions based on the data. Ultimately, it is a “system … that learns to generate more objects that look like the data it was trained on” (Zewe, 2023, para. 4). Much of what we see of these generative AI tools in academics are large language models, analyzing large amounts of text for patterns and connections between words and phrases and create text responses to prompts based on algorithmic probability of those patterns (Priest, 2024). In other words, it is a highly sophisticated predictive text model which doesn’t just predict the next word in your message but creates the whole message. This is based on the mathematical likelihood that one phrase will follow the next, given how other texts addressed the same topic. The tool also has the ability to improve and modify outputs
18 | P a g e
Made with FlippingBook - Online Brochure Maker