Focus points

How you deal with AI in your own teaching is, of course, not something you do on your own. It is therefore important to discuss this with your colleagues. Do you actually know what your fellow teachers' views are and how they deal with AI in their courses? Naturally, you want to impart certain knowledge about AI to your students, but this does not have to be done in the same way in every course. When in their study path are students asked to follow the student modules? When do you want to introduce them to the various aspects of responsible use of Generative AI? Which colleagues are best suited to do this? It is wise to tackle this together.

The Centre for Academic Teaching and Learning can provide support with these kinds of issues. For example, tailor-made workshops are offered to provide teaching staff with basic knowledge about Generative AI and to initiate and facilitate discussion about the different views on GenAI. This can be done during a staff meeting, for example. In addition, template presentations have been developed to help communicate with students about various topics. For questions and advice, please contact: generatieve-ai@uu.nl.

When it comes to summative assessment, it is not possible for individual teachers to assess the impact of Generative AI on education on their own. This must be addressed at the programme level, usually by the director of education. It is a good idea to discuss this within your teaching team, both for the course and the programme. Ultimately, the impact of Generative AI on the final attainment levels must be assessed for the entire programme. Do the current assessments still test what they are supposed to test? Are new learning objectives needed, or do existing learning objectives need to be adjusted?

Educational Development and Training has developed the test programme scan for teaching teams and education directors.

The AI policy provides clarity and frameworks in which situations AI can be embedded in our education.

• The AI responsibility matrix clarifies who in what context or situation is responsible for implementation, daily use and compliance with agreements around the use of AI within our education system.
• The AI index is a scale of five levels in the AI use by students.

Suspicion of fraud

Be careful if you want to ban generative AI. It is not (yet) possible to prove that a student has used a chatbot fraudulently. None of the current AI detectors are reliable, and it does not look like they will be anytime soon. This means they cannot be used as official evidence, because they generate too many false positives and false negatives.  

Furthermore, the detectors only check whether a specific chatbot has been used, not to what extent. This may be acceptable if students are not allowed to use generative AI at all, but in many cases, moderate use will be permitted. It is not clear whether a detector can identify whether a piece of writing has been written entirely by a chatbot or whether it has only been spell-checked.    

This makes it difficult to take measures retrospectively in cases of suspected fraud. If you suspect fraud, you can, of course, always talk to a student and/or the whole group.  

Therefore, if you want to ban generative AI, it is important to engage in dialogue with your students. By having an open conversation, you can involve the students in the considerations. You can try to establish a social contract with your students, as it were.

You can also adapt the assignment in such a way that it is less susceptible to fraud. You can ask your students to submit a kind of logbook of the creation of an assignment or, for example, first have them make an assignment outline that is then turned into a paper.

It is very important that you give clear instructions to your students. The greatest need among students in the field of Generative AI is transparency from teachers. They want it to be discussed openly and for clarity to be provided. Students indicate that they often do not know when something constitutes fraud and that they find it difficult to discuss this with their teachers.

So, for each assignment, indicate to what extent they may use Generative AI based on the AI index. It is particularly important that you explain why you have chosen a specific level of the index. Make clear what the learning objective of the assignment is and why a certain level of AI use is appropriate for it. As discussed in the previous point, fraud is very difficult to prove. Many students do not want to commit fraud, so clearly indicating what is and is not allowed (and why not) can help a lot. You can also make agreements with your students in this regard. Together, you can discuss what constitutes reasonable use of AI to achieve the learning objective and establish a kind of social contract.

In addition to communicating about AI use during the course, there are other topics that are important to discuss with your students. Of course, this does not all have to be covered in a single course and can be addressed at programme level, as mentioned in a previous point. To make this easier, we have created a number of template presentations. 

The University Library also offers various materials that students can use to educate themselves in the field of AI. These include AI E-modules, the libguide, and workshops for students. Students receive a badge after successfully completing the e-modules. This allows you to make it a compulsory part of your course. 

Do you want to give an assignment that deliberately incorporates generative AI? This can be a good way to encourage students to look critically at the output, learn what the limitations are and impart how to use generative AI responsibly. In any case, this can give you insight into how students use it.

One problem to consider here is that we cannot force students to use generative AI if a student does not want to because of privacy concerns. In that case, you can either offer the assignment as optional or offer an alternative assignment. Students can then, for example, ask someone close to them to help instead of using generative AI. That too can be interesting to reflect on and compare with students who used generative AI.

Research is being done by teachers in USO projects on how to do this in different ways. These results will be shared as soon as possible.

If you want to use generative AI in your teaching, think about how you can know afterwards whether your goals have been achieved. See more information on this under the following field on this page.

Not just your students, you as a teacher can also use generative AI for teaching tasks. Before you get started with this, it is important that you yourself understand what generative AI is, have made ethical considerations and know how to use generative AI effectively with prompting.

Some examples of teacher tasks that can be enriched with generative AI:

• Let generative AI act as a colleague: ask questions! What would you like to spar with an expert about?
• Get inspired with new forms of work, testing or lesson structures
• Redesign a course or lesson, for example with
  • The integration of educational principles
  • Applications of Blended Learning
  • Increasing inclusion and diversity
• Designing quizzes and (practice) test questions
• Generating assignments and assignments, including instructions
• Analysing patterns in test results including suggestions for improvement
• Developing a rubric based on learning objectives and criteria
• Generating standardised feedback
• Designing a tree structure for use in Brightspace

Above all, try out how generative AI can help you further develop your teaching. Could you use some help with this? Then email your question to generatieve-ai@uu.nl. For example, we can support you or your teaching team in effective prompt learning or guide and facilitate a discussion on the use of generative AI.

As in any Plan-Do-Act-Check cycle, it is important to reflect on how you have deployed generative AI for your teaching, so that you can learn from your own experiences (positive and negative). Hopefully, you can also share this with others and thus contribute to our knowledge about generative AI in education! 

As early as possible, try to think about the question: how will I know afterwards if I have achieved my goals? This way, reflection becomes an integral part of using generative AI in your teaching. For example, you can collect feedback from students and colleagues using these feedback tools (in Dutch). Do you want more insight on whether deploying generative AI enhances your students' learning? Then try reformulating your teaching problem into a researchable teaching question using the eight-step Utrecht Roadmap for Scholarship of Teaching and Learning.