Simon Coupland, De Montfort University
The Higher Education sector is rightly ploughing considerable resources into its response to the emergence of generative A.I. The question of whether a student wrote an essay themselves is a surprisingly complex one, which, in truth can only even be answered by the student themselves. A great deal of money and effort has gone into technical approaches which seek to detect submissions created using generative A.I. Some of these approaches have quite high detection rates, but also suffer with false positive rates meaning most institutions would be falsely accusing hundreds of students per year of committing academic offences. Another approach is to help students embrace generative A.I. as a tool and require them to describe how they use such tools when completing their coursework.
This paper provides a new perspective on this question, not by examining a student's written submission, but by how a piece of work was written and edited. Such an approach is presented as complementary to automatic A.I. detection and student declarations of A.I. use. The author has developed a piece of software which tracks the development of a document over time. Each time the student digitally touches their work the changes are recorded. An overview of this data is presented back to tutors through a web interface, integrated into the virtual learning environment The software gives an overview for a cohort, with the ability to delve more deeply into an individual submission. The result is an easily accessible interactive history of the submission during its development giving some kind of provenance to the document. This history of construction and editing, shows how a piece of written work has been crafted over time, providing useful evidence of academic practice.
Data on the points where students digitally touch their work can be also useful beyond questions of academic practice. The paper presents how such data can be used to inform formative feedback throughout a module, targeting students where it is most needed. Results from applying this approach to a final year computer programming module are given. These show an increase in pass rate, without general grade inflation.
Conference theme: Generative AI