10 insights for harnessing artificial intelligence in scientific research: supervisor perspectives

Abstract

Background Artificial intelligence (AI) transforms scientific research and publication by supporting all aspects of research from setting up hypothesis and research objectives to data analysis and manuscript preparation. Its widespread use in postgraduate research by students improves efficiency and feedback, but also raises concerns related to academic integrity, ethical behavior, algorithmic bias, hallucinated outputs, copyright issues, and mainly, preservation of students’ critical thinking skills. This article presents supervisor-oriented guidance for the responsible integration of AI into scientific research supervision. Methods We have undertaken narrative synthesis based on published manuscripts as well as our collective supervisory experience with AI-assisted research projects. Published literature evidence and practical experience were integrated to develop 10 insights for supervisors guiding research students who use AI tools during all aspects of their research. Results We have generated 10 practical insights for supervisors on guiding research students with ethical AI use: Establishing transparency and accountability; Setting expectations; Creating a customised AI use agreement; Maintenance of AI Use Log for Review; Reviewing AI log – The supervisor’s role; Addressing data privacy and ethical risks; Assessment of students’ critical thinking and conceptual understanding; Guide verification and validation of AI-generated content; Prepare students to justify AI-related decisions during post-research evaluation; and Institutional support for supervisor competency. Together, these recommendations emphasize the importance of human oversight, documentation and verification to distinguish AI assistance and original student scholarship. Conclusions AI can enhance postgraduate research and supervision when used as a transparent and ethically governed tool rather than as a shortcut. Supervisors have a central role in ensuring students’ ethical research practices evident by clear documentation of AI use, verification and transparency. A structured supervisory approach following the 10 insights can help balance the efficiency of AI tools with research integrity, accountability, and meaningful human mentorship.