The GSLS has updated its Generative AI (GenAI) guidelines for students, teachers, and for research and supervision, to reflect the GSLS’s adoption of the UU’s AI Index. To further support you, we have also updated our educational tutorials for teachers which can be found on the Ulearning platform (log in with your Solis-id). These guidelines and tutorials are aimed at enhancing your teaching practices while addressing the challenges and opportunities GenAI presents. The full teacher and supervisor guidelines can be found below. It is recommended to take a look at the guidelines for students as well.

The AI Index

The GSLS has adopted the UU AI Index, a five-level framework for how AI may be used in assessed work. As a teacher or supervisor, you decide which level applies to your course or assignment, and this must be clearly communicated in the assignment brief, course syllabus, and on Osiris/Brightspace. Students should never have to guess what is permitted or not. It is also worth explaining why a level applies, not just stating the number, so students engage with the boundary thoughtfully rather than simply trying to stay within it.

GSLS’s boundary is Level 3, unless you set a different level. We’ve set this as the boundary because building the field-specific knowledge and judgement to evaluate AI output underpins our master’s education and cannot be borrowed from a tool. Level 4 and 5 can be entirely legitimate in the right context, but they require sign-off precisely because they shift more of the intellectual work onto AI.

Across all levels, students retain the right to opt out of AI use unless AI skill-building is itself a learning objective. Where AI use is built into an assignment, an alternative non-AI pathway must be offered, or explicitly mentioned in the course description when this is not offered. Where AI is used in your course, it should be usable through free tools only, so no student is disadvantaged by cost.

UU also offers EduGenAI, a SURF-hosted platform for working with AI in a secure environment, free to teachers and students who request access via this activation form.  It must be noted that this tool is for educational use only and is not permitted for scientific research. UU is also developing its own AI educational platform, SolisChat, which will let teachers build course-specific chatbots that automatically enforce the correct AI Index level for that course; this is expected in the coming academic year. Once available, details will be added here.

For the full AI Index framework, detailed examples, and guidance on designing assignments around each level, see Module 4 of the GSLS GenAI Tutorials for teachers, and the UU AI Index teacher reference.

Level 3 of the UU AI index is applicable to the Major Research Project, Life Sciences Academy, Profiles, and the Writing Assignment. The student should always discuss with their teacher or supervisor before using AI.

Highlights teacher and supervisor guidelines

• Embrace GenAI with transparency: communicate openly about GenAI’s capabilities and ethical considerations, fostering an environment where students feel comfortable discussing GenAI use. Modelling this yourself, being open about where GenAI helped with your own course prep can make it easier for students to be honest about their own use.
• Evaluate and adapt assignments: not all classroom activity is equally educational, some genuinely build understanding, some just look like it. Identify where real learning happens in your assignments, and adjust tasks vulnerable to automation to protect it, focusing on critical thinking, argumentation, and deep content analysis. See Module 4 for the full framework behind this.
• Incorporate GenAI thoughtfully: decide which AI Index level best fits your assignment, and use GenAI for tasks like generating quiz questions, facilitating research ideas, and helping adapt rubrics.
• Request disclosure: ask all students to write a disclosure statement on their use of GenAI, including students who did not use it in their work. For research and business reports and writing assignments, this disclosure is part of the GSLS Rubrics. 

Open access tutorials

Our teacher and supervisor tutorials (Modules 1-4) cover the same foundational materials as the student tutorials, plus a dedicated module for your side of teaching with GenAI. Module 4 in particular:

• Designing your course and assessments so GenAI supports genuine learning rather than replacing it
• Choosing and communicating the right AI Index level for a given course or assignment
• Example disclosure agreements and how to adapt them to your course
• Recognising and responding to suspected misuse, including what not to do
• Using GenAI thoughtfully in your own teaching and admin work

For further details or feedback, contact GSLSGenAISupport@umcutrecht.nl.

Additional guidelines for Generative AI in research projects

These guidelines serve as a complement to the guidelines mentioned above, specifically tailored to the context of research projects.

Before using GenAI in research projects, students must discuss with their supervisor how much they think GenAI can be used. Where a project involves unpublished research, patient or patent data, using GenAI with that data requires prior approval from the supervisor, the relevant research office, and the ethics committee, regardless of the platform account used. Unauthorised use of GenAI can be considered fraud as specified in the EER.

GenAI research project guidelines for Master’s students and supervisors

Steps to take when you suspect inappropriate use of Gen AI tools

Detection tools can identify direct plagiarism reasonably well, but they are far less reliable for more common issues: a student working beyond the AI Index level set for their assignment. These tools can frequently flag false positives, and non-native English writers and neurodivergent students have been known to be flagged disproportionately. Treat any detector output as a prompt to look closer, never as proof.  

What’s worth noticing. Start by checking the disclosure statement against what you’re seeing. Beyond that, watch for a noticeable shift in style or quality from previous work, generic statements that don’t fit the student’s usual angle, inconsistent citations, or overly verbose writing. However, none of these are proof alone.

Start with a conversation, not an accusation. Ask the student to walk you through how the work developed and where they used GenAI. You can request their prompt history if you stated in your assignment that this could be collected. Most cases resolve themselves at this stage. If it’s a genuine misunderstanding, resubmission (of the affected parts, at the correct index level) or a short oral/closed-book check is often better than a penalty or escalating to the Examination Broad right away.

Never paste a student’s work into an AI tool to ‘see what it says’. It’s their personal data and intellectual work and not yours to upload, even informally. The only exception is an approved research project with ethics board approval and informed consent.

If real doubt remains, or a student won’t engage, escalate to the Board of Examiners for advice, rather than deciding alone.

See Module 4 (4.6) for the full guidance.