Description
As generative AI tools disrupt education at all levels, physics educators must adapt to a new and rapidly changing landscape. I present a review of current educational practices and literature on the benefits and risks of generative AI use in student learning in physics, mathematics, and coding.
I focus on the impacts of cognitive offloading, the process by which people can reduce the cognitive effort required for a task by using tools, which can range from simple (pen and paper notes or basic calculators) to complex (such as generative AI) [1].
Cognitive offloading can provide significant benefits by allowing a person to focus their thinking on the portion of a task that matters most. However, cognitive offloading also poses risks: inappropriate cognitive offloading can reduce learning and the quality of student work [2].
For generative AI tools in particular, there is research suggesting this cognitive offloading is beneficial for experts completing familiar tasks [3] while another emerging body of work highlights potential harm to novices learning new skills and content [4-7].
Drawing on this literature as well as data from surveys of students at Monash University, I share strategies for scaffolded adoption of AI tools in physics classrooms, suggestions for where AI-free learning may support physics students’ learning, and approaches to collecting internal data from your students to better understand their evolving AI use.
References
[1] E. F. Risko and S. J. Gilbert, Trends in cognitive sciences 20, 9 (2016).
[2] Y. Z. Fan et al, British Journal of Educational Technology 56, 2 (2024).
[3] K. Z. Cui et al. Management Science (2026).
[4] K. H. Fredly, et al. arXiv preprint arXiv:2603.06342 (2026).
[5] J. H. Shen and A. Tamkin. arXiv preprint arXiv:2601.20245 (2026).
[6] C. Liu, et al. arXiv preprint arXiv:2604.04721 (2026).
[7] S. Sankaranarayanan, arXiv preprint arXiv:2602.20206 (2026).
| I am the presenting author | Yes |
|---|