Investigations reveal hundreds of UK students using AI tools like ChatGPT in their assessments.
Recent figures obtained by The Tab reveal that over 40% of UK universities have caught students using AI chatbots, such as ChatGPT, to cheat in their assessments.
Responses to Freedom of Information requests from 114 universities show that since December 2022, at least 377 students from 48 institutions have been investigated for using AI chatbots in their university-assessed work. So far, at least 146 students have been found guilty, with many investigations still ongoing.
The actual number of students using AI chatbots to cheat on their exams could be higher, as some universities either don’t have centralized data or find it too costly to extract. One university with held data, citing concerns of potential reputational damage.
Investigations are difficult
The University of Kent holds the highest record, with 47 students under investigation for using ChatGPT or similar AI chatbots. Despite providing AI guidance, training webinars, and guidelines for students, the university found 22 guilty of plagiarizing AI-generated content in their assessments.
The investigations have taken a considerable amount of time, and most are still open due to the novel technology involved. Leeds Beckett University, which has the third-highest number of investigations (35), is “grappling with the rapidly evolving situation regarding generative AI tools.” So far, 10 students have been found guilty, and 19 investigations are ongoing. At Birkbeck, University of London, fewer than five of the 41 investigations have resulted in a guilty plea, with the majority still open.
Cheating has been found at universities across the rankings, from the University of Bedfordshire to the prestigious St Andrew’s. Students caught using AI chatbots have received penalties ranging from a zero grade to failing an entire unit, and being required to attend academic tutorials or workshops. You can find a table of all known investigations to date at The Tab.