Guidelines for teaching staff
KU Leuven’s view: "Don't ban GenAI, but use it responsibly"
KU Leuven has decided to promote a responsible use of Generative AI (GenAI) in education instead of banning it.
Apart from the general university principles, the following additional principles apply specifically for educational purposes:
- Teaching staff are expected to clearly inform students about whether or not they are allowed to use GenAI for assignments such as visual, writing and programming assignments.
- Students are expected to be transparent about the use of GenAI so that their knowledge, understanding and skills can be assessed fairly and correctly.
- If a student uses GenAI and is not transparent about its use, this may be considered an irregularity. (According to article 84 of the Education and Examination Regulations).
The guidelines on this webpage apply to the whole university. This means that the guidelines on this page apply only if the faculty, the study programme or the teaching staff member has not given any additional information.
What is GenAI? Basic principles
Quick links
GenAI and students
- Which information do you give students on the use of GenAI?
- When do students have to be transparent about the use of GenAI?
- How can you use GenAI?
- Code of conduct
GenAI and student assessment
- How to assess results?
- Detecting GenAI
- What do the Education and Examination Regulations (OER) say about the use of GenAI?
GenAI in the future
- How do we prepare for the potential use of GenAI in future academic years?
- Sources
See also:
Basic principles
- Guidelines for students
- Guidelines concerning research
- The use of chatbots in education
Students and GenAI
Which information do you give students on the use of GenAI?
Students are using AI. This is a reality we cannot ignore. However, we want them to use AI in a responsible and creative way and apply a critical approach. We want them to understand its capabilities and limitations.
Teaching staff should keep the following recommendations in mind:
- Share the following guidelines on the use of GenAI with your students:
- the 2 basic principles for students:
- The student has full responsibility for what they submit.
- The student makes sure that the assignment allows the you to evaluate which competences they have acquired as a student.
- Instructions to cite and reference GenAI;
- How to be transparent about GenAI;
- General principles, tips and tricks for responsible use of GenAI.
- the 2 basic principles for students:
- Discuss the specifics of your course:
- When a student gets an assignment, you clearly indicate what is and what isn't allowed.
- Clearly communicate the limitations of GenAI to your students.
- Discuss how you expect students to be transparent about the use of GenAI in their assignments, and how they have to report and monitor it.
- Give students information on how they have to cite and reference GenAI.
- If necessary, share the guidelines around the use of GenAI in researchwith your students.These are potentially relevant in master's theses or other course units that focus on research skills.
When should students be transparent about the use of GenAI?
A GenAI model can be used in different ways. Sometimes, the use of GenAI is similar to what we already do with other tools. In this case, we do not need additional transparency. However, other applications may make it more complicated to assess a student’s contributions to an assignment and require more transparency.
You might consider asking your students to add extra information on if and how they used AI in their assignment. They could, for instance, take screenshots of the entire exchange with GenAI, highlighting relevant information, or they could add an explanation of how or why GenAI was used.
Examination committees can also ask students to discuss the assignment they submitted. This will give the student a chance to be more transparent. After all, correct use of GenAI and being able to explain it is an essential part of acting with scientific integrity.
How can your student use GenAI?
The student is not allowed to use GenAI:
- Any form of copy-pasting of any content generated by GenAI without fully acknowledging the source (citing, referencing). After all, if you want to produce any academic work, you need to check the original sources. Using GenAI with proper referencing can only be done in exceptional situations.
- Any use of GenAI that prevents the teaching staff from assessing your skills.
For example: using paraphrasing tools on texts you did not write yourself to cover up plagiarism, misuse of translation software. - Using AI-generated content during on-campus exams or other evaluations when it has been communicated that the use of GenAI is not allowed.
The student is allowed to use GenAI without explicit reference:
- Using GenAI as a language assistant for reviewing or improving texts you wrote yourself, provided that the model does not add new content. In this case, the use of GenAI is similar to the spelling and grammar check tools we already have today, so you do not need to explicitly mention using GenAI for this. Again, te student always makes sure that the use of GenAI does not interfere with an examiner's evaluation of their achieved learning goals.
- Using GenAI as a search engine to get initial information on a topic or to make an initial search for existing research on the topic. This way of gathering information is similar to using an ordinary search engine when working on an assignment. Students are then expected to look for scientific sources and to interpret, analyse and process these source documents. If they then write their own text based on this information, they don't have to mention they used GenAI.
If, by way of exception, they do copy-paste certain parts of GenAI output (for instance because of the nature of the assignment), we expect them to cite their sources.
The student can use GenAI when explicitly allowed by teaching staff:
- Having GenAI generate code as part of a larger task.
- We expect students to consult with the coordinator of the assignment (teaching staff member, supervisor, assistant ...) if they want to generate audio or visual content with GenAI for the assignment so they have clear information on expected transparency and referencing.
GenAI and assessment
How can you be mindful of the use of GenAI in the assessment of results?
Teaching staff use the following 3 basic principles:
- You assess assignments (or other examination content) assuming the student personally created the content.
- If the student indicates where this is not the case, this is taken into account in the assessment (to what extent has the student achieved or not achieved the competences?).
- If the student passes something off as their own work or their own acquired competences when there are indications that this is not the case, is followed.Notify the chair of the examination board when an irregularity is detected.
Things to consider to get a better idea of the use of GenAI:
- Ask students to be transparent about the use of GenAI. You can refer to the Code of Conduct.
- For paper: add an additional test or a discussion during an oral exam, etc.
- For the master's thesis: ask targeted questions about parts that stand out (sudden change in style, lack of depth, lack of references, incorrect references, etc.) or questions about the creation process, etc.
Detecting GenAI
- The Education and Examination Regulations apply in full to use of GenAI, specifically arts 84, 85 and 86.
- Turnitin’s AI detector is no longer available since the start of the academic year 2023-2024. Numerous investigations in 2023 resulted in its unreliability. On top of that, the detector only worked for English texts and was limited to 15.000 words.
- We strongly discourage the use of AI detectors because of potential issues with privacy and GDPR guidelines. If you do wish to use an AI detector, it is only possible if you inform your students when instructing the assignment.
- Consider the AI scores of AI detectors as purely indicative. We urge caution given AI detectors can also generate false negatives and false positives.
- Look for clues other than AI scores. The result of an AI detector is not enough to establish an anomaly. Thus, even now, a similarity score of Turnitin is not a sufficiently conclusive argument to impose a sanction.
- Other possible indications are: a clear break in writing style, deviations in content, indications via the references, the student not being able to explain or defend the submitted paper ...
- If a confluence of circumstances or a combination of indications is established, which can no longer be explained from chance, this may well constitute sufficient grounds for imposing a sanction.
- Teachers or supervisors who suspect based on several of these indications are advised to consult with the faculty plagiarism expert (in Dutch) so that the appropriate steps are taken.
What do the Education and Examination Regulations (OER) say about the use of GenAI?
According to Article 84 of the Education and Examination Regulations (OER), any use of a tool that wasn't explicitly permitted is considered an irregularity.
‘Every conduct individual students display with which they (partially) inhibit or attempt to inhibit a correct judgement of their own knowledge, understanding and/or skills or those of other students, is considered an irregularity which may result in a suitable penalty.’
All irregularity have to be reported
to the chair of the examination committee.
Use of GenAI in the future
How do we prepare for the potential use of GenAI in future academic years?
Using GenAI brings both challenges and opportunities for assessing learning outcomes and teaching critical skills to students. It is important that you already start reflecting on whether and how GenAI can have a place in the study programme(s) you are involved in.
Talk about this in the POC, and thoroughly reflect on this with teaching staff and students:
- How does the use of GenAI relate to the learning goals and personal development goals of specific courses and the study programme?
- What is the added value of using GenAI and where does it have a place in learning sequences concerning information, writing or research skills, or in assignments with a critical or creative aspect?
- How do we teach students the limitations of GenAI, and how do we teach them to use it critically and responsibly?
- What frameworks and constraints should we develop so that students learn how to use GenAI responsibly?
- How do we make sure the acquisition of basic factual knowledge remains a key pillar for critical thinking, reasoning and seeking solutions to issues - and this despite the use of GenAI?
- How do we ensure that students acquire basic skills such as academic argumentation and writing, when they can also use GenAI for this purpose?
- How do we ensure that students see learning as something of value now that GenAI makes it so easy to generate a fast result?
- How do we adapt our assessment to this context?
Sources
- What if my students used a chatbot? - KU Leuven Learning Lab
- Inspired by ACL 2023 Policy on AI Writing Assistance - ACL 2023 (aclweb.org)
- APA referencing style-MLA referencing style-Chicago referencing style
- Alkaissi H, McFarlane S I (February 19, 2023).Artificial Hallucinations in ChatGPT: Implications in Scientific Writing. Cureus 15(2): e35179.
- Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021, March).On the Dangers of Stochastic Parrots: Can Language Models Be Too Big?. InProceedings of the 2021 ACM conference on fairness, accountability, and transparency(pp. 610-623).
- Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., ... & Amodei, D. (2020).Language models are few-shot learners. Advances in neural information processing systems, 33, 1877-1901.
- Farrokhnia, M., Banihashem, S. K., Noroozi, O., & Wals, A. (2023). A SWOT analysis of ChatGPT: Implications for educational practice and research. Innovations in Education and Teaching International, 1-15.
- Gozalo-Brizuela, R., & Garrido-Merchán, E. C. (2023). ChatGPT is not all you need. A state of the art review of large generative AI models. 22.
- Perry, N., Srivasta, M., Kumar; D., Boneh, D. (2022) Do users write more insecure code with AI assistants?
- Artificial Intelligence and the Research Paper: A Librarian’s Perspective – News (smu.edu)
- If we are setting assessments that a robot can complete, what does that say about our assessments? | by Daisy Christodoulou | The No More Marking Blog
Dalalah, D., & Dalalah, O. M. A. (2023). The false positives and false negatives of generative AI detection tools in education and academic research: The case of ChatGPT. The International Journal of Management Education, 21(2). https://doi.org/10.1016/j.ijme.2023.100822
Elkhatat, A. M., Elsaid, K., & Almeer, S. (2023). Evaluating the efficacy of AI content detection tools in differentiating between human and AI-generated text. International Journal for Educational Integrity, 19(1), 17.
Liang, W., Yuksekgonul, M., Mao, Y., Wu, E., & Zou, J. (2023). GPT detectors are biased against non-native English writers. Patterns (N Y), 4(7), 1-4. https://doi.org/10.1016/j.patter.2023.100779
Perkins, M., Roe, J., Postma, D., McGaughran, J., & Hickerson, D. (2023). Game of tones: Faculty detection of GPT-4 generated content in university assessments. ArXiv, 1-21.