School of Mathematics & Statistics

Generative AI and Assessment in the Mathematical Sciences

As our technological landscape rapidly evolves with the success of generative AI, so must our learning outcomes and assessment practice, to ensure we prepare our students for the modern working world. The SDAEG group have hosted an LMA/IMA/RSS-funded workshop to bring together educators from across the UK to discuss innovative assessment ideas, potential issues, and how to continue to pass on relevant knowledge. Assessment changes have been trialed across our online and in-person masters programmes, particularly in courses containing coding, a skill in which generative AI is well-versed. Innovative ideas include developing bespoke Python libraries which the students interact with, and including peer-review of coding via GitHub. A traffic light system has also been applied to one programming course to help guide students on generative AI usage. For example, class tests come with a red light (no AI usage), collaborative with AI assessments come with a green light (usage encouraged) and lab sheets come with an amber light (try to solve the problem yourself first, ask demonstrators for hints, and if still stuck, use generative AI).