A recent BBC report has highlighted the difficult questions universities face when assessing whether students have used generative artificial intelligence in their work.
Harrison Sharples, a 20-year-old medical student at the University of St Andrews, was accused of using AI to write a dissertation on play therapy and anxiety in hospitalised children. He maintained that he had written the work himself and had not considered using AI.
According to the BBC, one of two human markers raised concerns about the dissertation’s polished and uniform tone, consistent grammar, similar sentence structures, formulaic paragraphs, repetitive phrasing and use of em dashes. These were treated as possible indicators of AI-generated writing.
The case illustrates a central difficulty: characteristics commonly associated with AI are not necessarily proof of AI use. Clear grammar, a consistent structure and particular punctuation may simply reflect a student’s natural writing style or careful editing. Meanwhile, AI systems are developing rapidly, making their output increasingly difficult to distinguish from human work.
Following the concerns, Harrison attended a virtual academic misconduct hearing. The link arrived after the hearing had already begun, and he felt the disruption affected his ability to explain himself. The allegation was initially upheld, his mark was capped at the minimum pass level and the outcome could potentially have affected his application for professional registration.
He appealed. The university accepted that the delayed link had affected his opportunity to present his case and arranged a second hearing. After he was questioned about his writing process, the allegation was not upheld and he received his full dissertation mark.
Although university disciplinary proceedings are not criminal trials, their consequences can be significant. An adverse finding may affect a student’s results, graduation, future studies or entry into a regulated profession. A fair and transparent process is therefore essential. Students should be told clearly what is alleged, shown the basis for the concern and given a meaningful opportunity to respond.
Students can help protect themselves by retaining research notes, outlines, drafts and version histories. They should also check their institution’s AI policy and keep a record of any permitted tools they use. If challenged, this material may help demonstrate how an assignment developed.
Universities should also be cautious about relying on writing style or automated detection tools as proof. Concerns should be considered alongside the student’s working materials, subject knowledge and explanation of their methods.
Harrison was ultimately cleared, but the process caused him considerable stress, disrupted his studies and delayed his graduation. His experience shows why universities must get the procedure right from the outset.
Gibson Kerr can assist students involved in disputes with universities, including academic misconduct allegations, disciplinary proceedings, complaints and appeals. If you are facing a university dispute and would like advice about your position or the options available to you, please contact us for more information.
DISCLAIMER
The content of this page is for information only. It is not intended to be construed as legal advice and should not be treated as a substitute for specific advice. Gibson Kerr Ltd accepts no responsibility for the content of any third party website to which this webpage refers. Gibson Kerr Ltd is regulated by the Law Society of Scotland.