Using participant-facing AI safely in your research
12 Nov 2026
New guidance for researchers, postdocs, PGRs and research-related PS colleagues where AI is the subject of investigation, forms part of the research intervention or interacts directly with participants.
The University has produced guidance for colleagues on the safe use of g participant-facing AI- specifically where AI is the subject of investigation, forms part of the research intervention or interacts directly with research participants in a way that may affect their experience, behaviour, decisions or the data collected from them.
Examples include:
- AI chatbots
- Clinical decision-support systems
- AI interviewers to collect research data
The guidance gives colleagues information on ethical issues associated with the use of participant such as informed consent, participant welfare and safeguarding, accuracy and bias.
It also explains the process around using AI systems other than the two University-approved AI systems (Microsoft Copilot, Enterprise version only, and Nebula One).
If you're planning research involving participant-facing AI, please review the guidance as part of your study planning and ethics preparation.
Colleagues supporting research are also encouraged to familiarise themselves with the guidance and share it with relevant research teams.
Please note: The guidance does not apply to the use of AI solely as a productivity or research support tool (e.g., proofreading, literature searching, administrative tasks, or researcher-assisted data analysis), where participants do not interact with the AI system directly.
Such uses should instead be managed in accordance with the University's broader guidance on the responsible use of AI at: