Case Summaries

The panel considered a full transcript of the student’s viva. The panel decided that the student had committed two different academic misconduct offences. Firstly that their work included unacknowledged material generated by AI and secondly, that there was collusion with another individual.

The student was required to retake the module with uncapped marks. They were told that their overall degree mark would be lowered by 10% and they were also asked to submit a reflective essay on academic integrity.

The student appealed against both the finding of misconduct and the penalties applied. They were concerned about the reliability of AI plagiarism detection software. They said that the provider had also misunderstood the role of the external tutor they had consulted. The provider dismissed the appeal.

The student complained to us. We partly upheld the student’s complaint (we found it “Partly Justified”).

We decided that the provider had followed its academic misconduct procedures, and that its decision that collusion had taken place was reasonable. The evidence showed that the input the student had received from another person went beyond was what allowed in the provider’s proof-reading policy.

However, the provider had not shown what evidence led the panel to conclude that the student had used AI to the extent that they committed academic misconduct. We thought that this could have impacted the student’s ability to make an effective appeal.

We also thought that the academic misconduct panel misrepresented what the student had said during the viva. The provider said that during the viva the student admitted that they had used AI to translate and paraphrase texts. In the full transcript of the viva, we could see that the student said that they had used Google to find synonyms for words because English was not their first language. But the viva transcript and records from the panel hearing recorded no admission to unauthorised use of translation software or AI to paraphrase their work.

The provider’s procedures said that when misconduct is alleged, it would ask to see notes, draft plans, and earlier versions of work so that students can demonstrate how they developed their work.  It did not consider this kind of material in this instance. It also did not consider whether Turnitin's AI detection might be less reliable for non-native English speakers, which was relevant given the student's international status.

We recommended that the provider reconsider the misconduct allegations afresh, and allow the student the opportunity to submit a new supporting statement. We asked the provider to outline the specific offences under consideration and provide the student with all information relevant to its review.