Russell Group universities expel students for AI-assisted cheating
Four UK Russell Group universities expelled students for using generative AI to cheat on submitted academic work in 2024-25, with over 2,000 AI-related punishments recorded across the sample group—nearly triple the previous year. Academics suggest the actual cheating rate may be substantially higher, with only the most obvious cases caught before expulsion.
What happened
Four Russell Group universities—University College London, Imperial College London, University of Glasgow, and University of Leeds—expelled students for using generative AI in their submitted work during the 2024-25 academic year.
Across a sample of 350,000 Russell Group students, there were 2,053 recorded punishments for AI use in 2024-25, compared to around 700 in 2023-24, representing an approximate rate of one in 170 students.
Reported impact
A Russell Group professor stated that while most students were likely using AI in some capacity, only the most egregious cases were detected and punished. Evidence of undetected AI use included formatting errors, changed quotation marks, different fonts, and incorrect referencing.
The rise in detected cheating has prompted universities to reconsider academic integrity practices, with concerns that students are modifying their writing styles to avoid detection by AI-flagging software, including avoiding em dashes.
Index analysis
A Higher Education Policy Institute survey found that 12 per cent of British students admitted copying AI-generated text directly into assessments, up from 3 per cent in 2023. Seven of the 24 Russell Group universities reported they do not record AI-related disciplinary investigations.
The Russell Group director of policy stated that expelled students represented a very small proportion of the student population, though academics suggested the actual cheating rate may be significantly higher than detected due to AI sophistication.
Governance controls
Many universities use software such as Turnitin for assessed work to detect AI use, but this software has been frequently criticised for flagging legitimately student-written work.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.