British universities are moving back to the facet, while 95% of students report using AI in their studies compared to 66% in 2024.
The use of AI among British students has ceased to be marginal: 95% say they use AI in their studies, compared with 66% in 2024, and the proportion of those who directly insert text generated in noted work has increased from 3% to 12% over the same period. In the face of these figures, several institutions return to the examination monitored in the room.
The Birkbeck Law School has re-established supervised tests this year, and the result has surprised the teacher: the share of very good mentions has increased from 13% last year, with essays to be given from home, to more than 45% with supervised tests. Professor Stewart Motha, speaking of a staggering result, notes that the current presence and commitment have progressed in parallel. Swansea predicts a similar movement, Bath and Cardiff multiply the supervised formats. The University of Bath adopts a two-way approach for 2026-27: open tests, where AI is optional or even expected because it corresponds to professional practices, and closed tests, limited in time, monitored and face-to-face.
Other establishments — Bristol, Leeds, Nottingham — prefer a tricolor system of lights, from red to green that encourages. The substantive debate remains open: for Edinburgh Napier's Sam Illingworth, there is no evidence that students cheat more than before, and the university would gain from trust. A HEPI report notes that of 96 universities whose policy is searchable, 83 use a teaching vocabulary, but there is a systematic gap between what the texts announce and what they actually organize.
Read the full article on the original source
Go to the source →