Leading UK universities are struggling to detect and penalize students using artificial intelligence (AI) to cheat, despite widespread use of AI tools like ChatGPT for academic purposes. A new report reveals that fewer than one in 400 undergraduates at research-led Russell Group universities were punished for AI misuse last year, even though over 90% of students admit to using AI large language models in their studies.
A survey by the Higher Education Policy Institute (HEPI) found that nearly a fifth of students admitted to copying directly from AI chatbots. Josh Freeman, policy manager at Hepi, commented, “Given that almost all students use AI to help with their assessments, this data suggests a strikingly low proportion of students are being penalised for AI misuse. It is very likely that great numbers of students are slipping under the radar.”
Despite a two-year-old pledge by Russell Group universities to uphold “academic rigour and integrity,” nine of the 24 universities surveyed reported they did not track AI-related sanctions. Of those that did, an average of 74 students were investigated for AI misuse, with 51 facing punishment, a small number considering the universities’ total student populations.
Experts suggest that AI-enabled cheating is much more widespread than the data shows. A recent study by publisher Wiley found that nearly half of university students believe generative AI tools make cheating easier. Furthermore, Hepi found that 8% of students admitted to using chatbots for assessments without altering the output.
The use of AI in academia has prompted mixed responses. While some universities allow limited use, such as for creating lecture notes, experts argue that AI’s role in cheating is significantly higher. One student, who anonymously shared their experience, admitted to using ChatGPT extensively for coursework, sometimes copy-pasting the output without detection.
Course leaders from Russell Group universities also reported concerns, with one estimating that at least a third of students at their institution violated academic integrity rules using AI tools. Four universities—Durham, King’s College London, Leeds, and Queen Mary University—confirmed expelling students for AI misuse.
The Russell Group acknowledged the challenge posed by generative AI tools, stating that universities are working proactively to develop policies that ensure ethical use of AI while maintaining academic integrity. However, critics argue that the current measures are insufficient to address the scale of the problem.
