AI Principles

AI Strategy – Expanding Thinking, Not Replacing It.

Artificial intelligence is now part of everyday life, even or especially at universities. You probably already use one or the other AI tool, and that's okay. Anyone who completes a (business) degree today must be able to use these tools. Used correctly, they can support learning in a meaningful way, for example through additional explanations, exercises or feedback. However, if used incorrectly, AI can also impair the learning process if it replaces your independent thinking and problem-solving. However, it is precisely these phases that are crucial for understanding connections and building knowledge in the long term.

We do not want to ban the use of AI nor promote its unrestricted use. Rather, we want to convey a conscious and reflective approach to AI.

Depending on the learning objective, AI can be very helpful or make the actual learning success more difficult. The following principles are intended to provide orientation as to when and how AI can meaningfully support the learning process in our understanding. They are based on the learning objectives of courses as well as on findings from current teaching and learning research, referring also to the official AI-principles of WU as well as the Code of Conduct.

1. Think first, then use AI. The first few minutes of a new task should belong to your own brain. Think about the required approach yourself, develop a first draft of the problem and the solution yourself. Mistakes are okay, they can and will happen. The learning effect arises from creating and discarding your own approaches.

2. AI replaces work, not thinking. Only delegate what you understand yourself. Use AI for routine, formulation, structure and feedback, not for actual understanding. Outsourcing uncontrolled things prevents the very build-up of knowledge that is the focus of all courses. For example, use AI to have a solution to a problem that you have developed critically questioned, or let AI create exercise examples for you on individual topics, which you then solve.

3. Difficulty is a learning signal. If a task feels exhausting, then you might actually learn. The mental "struggle" is not an obstacle, but part of the learning process. But distinguish between two types of difficulties: Productive: You have understood the question but the path to the solution is still unclear? Then keep trying yourself, because then you are learning something! Afterwards, feel free to get feedback on your solution approach from the AI. Unproductive: You don't understand the question or task even after reading it for the third time or you lack the prerequisites? Then get support from colleagues, faculty or even from an AI.

4. Understanding beats speed. What you can't reproduce without AI, you haven't learned. A solution generated by AI in 5 minutes is worthless if you can't explain it yourself, apply it or transfer it to another situation later.

5. You take responsibility for your results. Regardless of whether you use AI or not, you are responsible for all submissions. Therefore, always critically examine results for their technical accuracy, plausibility and comprehensibility. AI can make mistakes, provide incomplete solutions or draw wrong conclusions. You must therefore be able to understand, comprehend and explain all your submissions yourself.