inary numbers shown on a range of computational devices
Artificial Intelligence

Cheats Never Prosper Edit

Dr Simon Coupland, De Montfort University 

Our newsfeeds are currently filled with stories about how generative AI technologies such as ChatGPT will take our jobs (Greenhouse, 2023) and render higher education largely pointless (Shearing and McCallum, 2023). It is certainly true that generative AI can produce the kind work we ask our students for whether it’s writing an essay on a given topic and synthesizing a piece of code in a specific computer programming language.  So how can we ensure our students are engaged in active learning and not just asking AI for all the answers? Perhaps technology already has the answer in the form of version control metadata. 

Metadata comes in many forms but broadly it is data about an artifact. We are used to cloud-based platforms such Google Drive or OneDrive allowing us to work collaboratively on documents. It is the rigorous bookkeeping of these documents and their metadata that make this possible.  This metadata provides a history of a document and documents with histories have not been copied and pasted. As an example, here is a graphical overview of the metadata of this article recorded on OneDrive.  I have measured words additions, words removed and overall file size.  These are plotted against versions; each version does have a time and date stamp, which I have not included here. 

The diagram shows the growth in the document size, and the clear editing of the document with additions and removals.  This trail of metadata shows the writing and editing of this short article. It would be possible to ask ChatGPT to write an essay and fake a trail of edits over time, however the effort involved is likely to be beyond that of the typical plagiarist. 

The uses for the history of a digital artifact go beyond spotting plagiarism, we can use this information to support students along their learning journey.  I teach programming and have long used GitHub (think Google Docs for code) for student coding work and submissions.  I use metadata from GitHub to monitor student engagement, progress and to provide timely, data-driven formative feedback.  This has seen an improvement in progression and attainment without top-end grade inflation (Coupland et al. 2023). 

Cleary, document metadata from version control systems can be useful for academics.  Right now, the challenge is accessing this data and presenting it in a straightforward, digestible format.  What is needed is a suite of easy-to-use tools that combine the setting of student work with the monitoring and analysing of versions and metadata. Without approaches like this this it will be very difficult to guarantee the integrity of student work in the future. 


Coupland, S.; Fahy, C.; Stuart, G.; Allman, Z. 2023. Real Time Monitoring of Student Progress for Formative Feedback and Interventions. In Formative Assessment and Feedback in Post-Digital Environments: Disciplinary Case Studies in Higher Education, Eds.  Elkington, S. and Irons, A. Routledge. Accepted for Publication. 

Greenhouse, S. 2023.  US experts warn AI likely to kill off jobs – and widen wealth inequality. The Guardian.  Available at: 

Shearing, H., and McCallum, S. 2023. ChatGPT: Can students pass using AI tools at university? Available at: 

Image: Image by Gerd Altmann from Pixabay 

Contact: To contact the author of this article please use simonc[@]