Director of Education statement on use of artificial intelligence tools in written work
Dear PGT and PGR students,
The ethos of the Department of Oncology postgraduate teaching programme is to develop critical thinking, analytical skills, research confidence and overall autonomy in discussing and defending a scientific hypothesis. Artificial Intelligence serves as a vital tool that can support this through brainstorming, outlining concepts, language refinement, or even simply formatting. However, excessive use of AI-support tools has a consequence on learning – especially in postgraduate education. Writing at the postgraduate level refines skills in analysis, synthesis, argumentation, and original thinking. Overuse of AI bypasses these processes, leading to shallow learning and risks failing to attain the cognitive skills our courses are specifically designed to develop. Critically, using AI makes research skills such as evaluating sources, building arguments, and structuring complex ideas obsolete – the very skills that are essential for your future academic or professional work. The greatest impact on students could be loss of confidence in their own ability to write and think. The consequence may be a weaker ability to explain or defend their own arguments, which would reveal to assessors a poor mastery of their subject, be it at Masters level or Doctor of Philosophy, which may be exposed at viva.
Furthermore, remember that AI is not perfect, especially around information quality and accuracy; for example, current AI-tools can hallucinate sources, misinterpret complex ideas, or generate outdated or fabricated references. This could lead to factual inaccuracies, again weakening your academic credibility. Notably, AI-generated content can appear coherent but might be considered superficial by examiners - lacking the depth, nuance, and original thought expected at postgraduate level. Our position is that while using AI is an important component, we define that the risk to academic integrity lies in content substitution, i.e., AI-generated content is original in wording but it is not original in thought - therefore could be considered as ranging from poor academic practice to "contract cheating" depending on the extent. For your future, as Postgraduates you are trained to produce independent thought and research - skills critical for MSc aiming for D.Phil. studies, or for D.Phil. students continuing in research, medicine, industry, or teaching - AI overreliance can leave students ill-prepared for these paths.
We carefully select all our students for those that will excel at Oxford and on our courses. You were chosen as you have the potential to be cancer research leaders of the future, and we encourage you take the opportunity to optimise your learning with us.
Eric O’Neill
Director of Education