Graduate School of Social and Behavioural Sciences
Use of GenAI
The use of generative AI and AI assisted technologies are more and more widely used. Neglecting the use of AI for PhD trajectories will be very difficult, since the tools develop further and get more advanced.
It is important for the PhD candidate to make conscious decisions about using gen AI in their research. We therefore provide the following guidelines:
Generally accepted, but good to discuss with your supervisor:
- language check/editing;
- code check.
Discuss with your supervisor:
- text generation;
- direct interaction between AI and participants to your research (Chatbot);
- data analysis and visualisation;
- producing scientific insights;
- drawing scientific conclusions;
- providing recommendations;
- generating theories or ideas;
- citing AI.
Discuss with data-steward and/or privacy officer:
- ALWAYS consult the data steward and/or privacy officer before developing a direct interaction between participants to your research and an AI; sharing data with an AI tool, for example for summarizing results and data-analysis. Some usages are prohibited by law, some others can only be carried legally if specific measures are taken. Situations that carry an additional risk are, but do not limit to, the processing (collection, analysis) of large scale data; data containing personal or sensitive information like: name, date of birth, medical history.
Points to take into account
Points to take into account
The final written text of the PhD thesis must reflect the intellectual contributions and critical engagement of the PhD candidate. While co-writing, editing, or assistance—whether by individuals or technological tools (including AI-based systems)—may support the writing process, such assistance does not diminish or transfer the candidate’s sole accountability for the content, interpretation, and integrity of the final text. Nonetheless, the candidate remains fully responsible for the scholarly quality, originality, as well as legal and ethical compliance of the submitted work.
The application of the technology should be under human supervision and control, and all work should be carefully reviewed and edited, because AI can produce authoritative-sounding results that may be incorrect, incomplete or biased. The PhD candidate is ultimately responsible and accountable for the content of the dissertation.
Given the limitations of generative AI, PhD candidates should only use generative AI when they understand the implications of doing so. PhD candidates should discuss the benefits and pitfalls of the use of generative AI with their (co-)supervisor and subsequently submit their research design to the relevant ethics committee.
All usage of generative AI and AI assisted technologies must de declared as such. If the PhD candidate uses generative AI or similar tools, tools, they need to be transparent about it in their publications and dissertation e.g. include in methodology section.