Speaker
Description
The rapid expansion of remote and technology-mediated assessment during and after the COVID-19 period exposed serious challenges to academic integrity, particularly in contexts where students were left to self-regulate with limited institutional oversight. These challenges have been amplified by increasingly easy access to generative AI tools, through which answers to problem sets and even full project reports can now be produced by simply uploading a photograph of an assignment to a large language model. In many university settings, AI-assisted cheating has spread rapidly through peer networks in ways that resemble social contagion \cite{sooknanan2017}.
Inspired by this trend, we present a differential-equation model that treats AI-assisted cheating as a social contagion driven by peer influence and moderated by institutional intervention. The model provides an accessible yet mathematically rich example for undergraduate teaching, allowing students to apply equilibrium analysis, local stability classification, and numerical simulation using a reduced planar system. A worked MATLAB example illustrates how analytical results are reinforced through phase-plane and time-series plots. By enabling students to explore academic integrity through quantitative modelling rather than moral instruction, this contribution offers a powerful and transferable approach for teaching mathematical epidemiology, differential equations, and applied modelling.
Bibliography
@article{sooknanan2017,
title = {When behaviour turns contagious: the use of deterministic epidemiological models in modeling social contagion phenomena},
author = {Sooknanan, Joanna and Comissiong, Donna M. G.},
journal = {International Journal of Dynamics and Control},
volume = {5},
pages = {1046--1050},
year = {2017},
doi = {10.1007/s40435-016-0271-9}
}
