The Practice
This case study is part of a BILT-funded project, AIvsAI, which investigates student experience and perspectives regarding process and procedures of AI-related integrity investigations across the Faculty of Arts, Law and Social Science. With additional funding from the School of Sociology, Politics, and International studies, further data from the case study broadened the scope of the original research that is based on 178 survey and interview participants at Bristol. Through the case study, we have reached an additional 207 students and 8 academics from four Chinese Universities.
This research contributes to evaluative and student-centred knowledge of policy and procedures regarding academic integrity investigations. This has led to better clarity in terms of making plagiarism referrals, as well as students and staff facing training and workshops across the Faculty. Additionally, it raises wider concerns in the field of AI and education, for which we will contribute an academic article to enrich the debate.
Findings
Social science students in Chinese Universities, including a couple of international students there, are more receptive to use generative AI in their studies. A higher percentage (96%) of them reported to have always or often used it; by comparison, only 16% of Bristol’s ALSS students reported to use AI frequently, and 41% used occasionally.
Whereas Bristol students reported confusion, especially those whose work has been referred for academic integrity investigations, about how and why their work was referred, and in some cases, dismissed, students from China were dealing with magic numbers. Be it 20% (oftentimes) or 15% similarity allowance in submitted work, students in Chinese universities were relatively fluent in lowering down their similarity scores, sometimes with the help of AI tools, ironically.
Nevertheless, from both sides, there have been cases where students’ work has been wrongly accused of plagiarism due to unreliable AI detection software indicators and the vows of this industry aiming to sell their products to students and universities. For example, in cases from Chinese universities, where individual students were instructed to write content in class – as that in in-person exams, such writing had been flagged as AI generated, forcing students to rewrite – sometimes in ridiculous ways, in order to submit their work to the grading system.
Moreover, adhering to the presumption of innocence approach, Academic Integrity Officers at Bristol had to dismiss some cases with very high AI similarity scores. In other words, even if a coursework is flagged as 100% AI generated, this number alone does not incriminate a student of plagiarism, at least not so in the faculty of Arts, Law and Social Science – as explained by Dr Aslak-Antti Oksanen at a SPAIS staff training workshop, who has held the Academic Integrity Officer post since 2021.
Although an experienced Academic Integrity Officer can identify telling signs in coursework that suggest an overreliance on AI, potentially with little substance as student’s own work, this will still be followed with further investigations – often through requests of drafts, examination of style and authorship, informal chats with the student, and /or a panel discussion that the student is invited to attend. Each step requires care and commitment from academic integrity officers and administrative teams.
Does AI challenge our existing principles of academic integrity? Yes, certainly. In the above examples, generative AI – with its power to generate content and its claim to be detecting and discerning its own words – has posed layers of intelligence challenges for us. Do either high or low AI similarity scores mean plagiarism? No, and yet, all AI cases are currently treated as potential “contract cheating” – a term that students rarely understood but got shocked by; basically, it means you contract and pay someone else to work for you and submit such work as yours, the worse kind of academic plagiarism.
Our attention to academic integrity, should, however, not blind us from another finding from the comparative scope of this research. While relatively less receptive social science students at Bristol had also expressed collective interest in better understanding how AI is used by their peers, by their lecturers and tutors, and by Academic Integrity Officers, their peers in China with seemingly deeper working relations with AI, raised acute concerns about the future of AI and their disciplines.
These concerns exceed the scope of this research, but the eagerness with which the students raised them, in open dialogues with their lecturers and professors and in our workshops in China, stayed with us. With the fast pace that AI tools get integrated into the education system, questions about AI and the future of various disciplines will soon confront us, higher education educators and practitioners. Imagine a scenario where students receive no constraints for using AI, how long would it take for our students to notice the soullessness that AI chats typically feature? And how long would it take for them to notice the gap between the knowledge and understanding they are assessed on, and the skills that they must acquire to succeed in their future careers? And, that gap, is what we must address in the coming era where AI’s presence in education is only likely to grow.
The Impact
This research worked with students whose coursework had been investigated. Together with a focus on social science students, this research provides evidenced discussions regarding students and staff insights, attitudes and perceptions held in the early stage of generative AI development and its use in higher education. It has contributed to better understanding and improved practices via school-based training, faculty discussions on academic integrity process and procedures, and global conversations on AI and academic integrity.
Next Steps
We are writing up a journal article to enrich the debate on AI and Academic integrity in higher education globally. Our findings will also continue to inform academic integrity policy and practices at the school and the faculty at Bristol through the University’s Task & Finish group on updating the contract cheating procedures.
Contact
Dr Lin Ma, Honorary Lecturer/ ALSS Study Skills Tutor
Dr Aslak-Antti Oksanen, Lecturer in Conflict, Security and Justice
Leave a Reply