Our understanding of the effects of GenAI use in education is still in development, and the topic also encompasses normative questions and differences between disciplines that make consensus (beyond certain basic principles) within the higher education community, or even our university or faculty, infeasible and undesirable. FGGA therefore places emphasis on the trust in teacher's professional judgement, which means that teachers and study programmes retain the freedom (and responsibility) to determine the role of GenAI in their courses and curricula, within the framework of the principles outlined in An FGGA approach to GenAI in Education. 

Practically, this means that teacher's GenAI attitudes and skills (beyond basic understanding) are of secondary importance to their pedagogical duties and expertise. The essential question remains the same: how do you design your learning activities and assessments so that they align with the learning goals for your course (and the programme's end terms)? Answering this question has undeniably become more complicated with the emergence of GenAI. Aspects of the answer might go beyond the scope of a course, and need to be tackled on the programme or university level. However, it remains a necessity for each teacher to take students's potential use of GenAI tools into account when preparing their course and designing their assessments.

GenAI use and fraud

Fraud involving GenAI can take many different forms, which can make it difficult to determine. The FGGA student guidelines state in general terms that any form of literal copying of GenAI-generated material without full source attribution, and any use without appropriate disclosure (as prescribed by the teacher), is considered fraud. In addition, more specific policies may apply per programme or course.

Clear examples of fraud through GenAI would be:

• The use of non-existent concepts or references in an assignment (the use of  fictitious research data, graphs, literature and literature references is usually considered fraud according to the Rules and Regulations, whether this is caused by GenAI use or not).
• Arguments that appear to be based on academic sources but that do not completely align with the information from the source (this can be difficult to notice, as it would require extensive knowledge of the academic literature or manual checks).
• Text that is directly copy and pasted from GenAI tools without proper referencing.

When a suspicion of fraud is referred to the Board of Examiners, the Board will conduct an investigation. The student will be invited for a hearing and the evidence will be examined, after which an official decision is taken. The disciplinary measures that can be taken by the Board of Examiners are determined by the Rules and Regulations of the programme. When fraud cannot be determined by the Board of Examiners, the grade for the assignment is released. The inconsistencies in the assignment can in that case be addressed in the grading if the issues are mentioned in the grading rubric of the assignment (for example mistakes in referencing, unclear language, etc).

In case of doubt, the examiner can always contact the Board of Examiners for guidance. They can provide more information on the possibilities and potential consequences.

Information and support

Below, you can find a continuously updated overview of resources on offer to support you in your teaching practice with regard to this topic. To start off, some general points to take into consideration (please be aware that your institute or programme might have more specific guidelines and policies in place!): 

• Clear communication about rules and expectations and your pedagogical justification for them with students is essential. It is advisable to include your policy in the syllabus and the assignment(s), as well as discuss it with students in class. The student guidelines include a Scale of permitted AI use that can be helpful for this. 
• Take a structural rather than merely discursive approach. This means avoiding reliance on student compliance with unenforceable rules (e.g. a take home essay with the stipulation that no AI use is permitted) and instead focusing on assessment design that inherently influences or constrains how students can interact with GenAI (see Corbin et al., 2025).
• A structural approach does not necessarily beget a mass return to classical supervised sit-in and oral exams. Always taking your learning goals as a starting point, try to imagine what evaluating skills and knowledge in an AI-rich environment would look like. Some tips for unsupervised assignments include:
  • Make them specific - involve in-class activities and discussions, current events and cases and student's own experiences and reflections.
  • Put greater emphasis on the process through scaffolding, peer-reviews and in-class components.
  • Encourage deeper student engagement by prioritising critical thinking, personal reflection, creativity and originality.
  • Complement GenAI-sensitive assessments with assessment formats less susceptible to unauthorised use of GenAI. You can also consider explicitly balancing formative 'AI-conscious' assessment with summative 'AI-proof' assessment (see also: Two-Lane Approach).
• Familiarize yourself with the basics of GenAI tools and test your assignments with them specifically in order to get a clearer picture of what GenAI use by students would look like for your assessment.
• Be aware that detecting and sanctioning student's unauthorized AI use is complicated: AI detectors are not reliable (and it is not allowed to opload student's work). This is why rethinking assessment design in advance is so important.

Individual consultation

Not sure where to start, feeling overwhelmed, looking for inspiration, ideas or a second opinion? We are happy to think with you! Send us your syllabus, assignment, or schedule a meeting right away. Contact Praneet Khandal and Eden Lutz via p.khandal@fgga.leidenuniv.nl and e.lutz@fgga.leidenuniv.nl.