Abstract

The integration of Artificial Intelligence (AI) in doctoral supervision within the field of Statistics has brought both opportunities and challenges for supervisors and PhD students. AI tools can enhance research efficiency, providing students with powerful means to analyse complex datasets, generate novel insights, and even draft parts of their thesis. However, the growing reliance on AI has also introduced new challenges in the student-supervisor dynamic, particularly concerning the trust and transparency in the research process. Both experienced and novice supervisors may find it challenging to assess the originality of a student’s work, as AI can mask the true extent of the student's understanding and contribution. For students, the temptation to over-rely on AI tools may hinder the development of critical thinking and problem-solving skills, which are essential for independent research.

The trust in the student-supervisor relationship is at risk of being eroded if students do not transparently communicate their use of AI in their research process. The capacity of AI to generate results or assist in writing raises concerns about the authenticity of the student’s work and the integrity of the PhD degree.

This presentation further discusses the strengths and limitations of AI tools in postgraduate Statistics research, drawing on feedback from surveys conducted among Statistics academics and postgraduate students at South African institutions. It also introduces a guiding rubric developed by the authors, which provides a set of suggested guidelines for supervisors and PhD students throughout the PhD process. While one aspect of the rubric focuses on maintaining transparency and trust through the responsible use of AI, its broader purpose is to enhance the overall supervision experience and foster the development of high-quality PhD students. The rubric outlines best practices for supervision, including the ethical use of AI tools as a complement to independent work and the inclusion of AI usage declarations in thesis submissions. The goal is to support the integrity of PhD research while optimising the potential of AI in academia.