I recently posted this scale (see below) on LinkedIn with the following text:
Where do you draw the AI line when it comes to student writing?
I reckon that for most of us it’s around 3-4, maybe 5-6 at a push, but I may be wrong!
A few people commented, including this:
“So much depends on what the output will be used for.
Are they writing guidance for a system that’s not safety critical?
Or a risk assessment for a complex and risky procedure?
And what – if anything – is being assessed? Because if its “how well can you use technology to quickly generate good enough guidance for your peers” we might be up in the high teens or even low 20s.”
Discuss?

Compiled from various online sources and resources including:-
AI Assessment scale (AIAS) (2026) Perkins, Furze, Roe and MacVaugh
Salatino, A., Fornari, F., Zdravkovic, J. (2025) CEUR-WS GenAI Usage Taxonomy https://ceur-ws.org/GenAl/Taxonomy.html
Ucan, S., Demirel Ucan, A. (2026) Generative AI in Student Writing: A Six-Level Taxonomy for Responsible Use in Higher Education Assessment. J Acad Ethics 24, 63
Assisted by GenAI (ChatGPT)
