*Thisarticle is based on a YouTube video by Sophie Hundertmark, an expert in the use of artificial intelligence with a focus on chatbots and strategic AI applications in companies and public institutions. Sophie is a researcher and lecturer at the Lucerne University of Applied Sciences and Arts and is doing her doctorate in Conversational AI at the University of Fribourg. The blog text was created using a custom GPT model that was trained on Sophie’s video content, language style and expertise. The result is well-founded, up-to-date articles based on Sophie Hundertmark‘s own expertise.
You can find the link to the video at the end of this article.
ChatGPT, Gemini, Kimi, and other AI tools have long since become part of everyday learning. Students use them to research information, write texts, create summaries, or find solutions. This presents a new challenge for teachers and parents: How can we design assignments so that students are allowed to use AI but still have to think for themselves?
A complete ban on AI is not necessarily the best solution. A much more interesting question is how we can adapt existing tasks to the new reality.
That’s exactly why I’m showing you three prompts that you can copy right away and try out with your own assignments. They can be used to rephrase traditional school assignments and homework in a way that places greater emphasis on the individual thought process, the application of concepts to one’s own examples, and genuine understanding.
Why We Need to Rethink School Assignments in the Age of AI
Let’s take a classic exercise from a German class:
“Write a summary of the first three chapters of *Wilhelm Tell* and analyze the first chapter.”
In the past, students had to read the material themselves, extract the key points, organize them, and put them into words. Today, an AI system can generate a perfectly usable summary and interpretation in a very short amount of time.
However, that doesn’t mean homework will become unnecessary. Instead, we need to change the problem statement.
Instead of evaluating only the final result, we can focus more on asking:
- How did the students arrive at their conclusion?
- What decisions did they make on their own?
- What was AI used for?
- Can you apply what you’ve learned to an example of your own?
- Can you explain this topic to someone else in a way they can understand?
The following three prompts address precisely these points.
Prompt 1: Make Your Own Thought Process Visible
In the first approach, the focus is not solely on the technically correct solution. Students must also explain how they arrived at their solution.
The use of AI is expressly permitted in this context. So their use is not hidden or prohibited, but is deliberately made an integral part of the task.
You can copy this prompt directly:
Du bist eine KI, die Hausaufgaben so umgestaltet,
dass der Denkprozess der Schüler:innen im Mittelpunkt steht.
Ich gebe dir eine bestehende Aufgabe.
Erstelle daraus eine neue Aufgabe, bei der:
- die Schüler:innen KI benutzen dürfen
- sie aber ihren eigenen Lösungsweg erklären müssen
Die Aufgabe muss enthalten:
- eine fachliche Hauptaufgabe
- eine verpflichtende Erklärung des Vorgehens:
- Wie bist du vorgegangen?
- Wobei hat dir KI geholfen?
- Wo musstest du selbst entscheiden?
Formuliere die Aufgabe so,
dass eine reine KI-Antwort nicht ausreicht,
um alle Teile zu erfüllen.Then simply add to your existing assignment.
The key difference: Even if an AI solves part of a technical task, the work is not yet done. Students must reflect on their work process and demonstrate which decisions they made on their own.
This shifts the focus from the result alone to the process of learning and thinking.
Prompt 2: Apply knowledge to specific examples
The second prompt takes a different approach. The goal here is to apply your technical knowledge to a real-world example.
The example should be drawn from everyday life, the students’ imagination, or their personal lives. This integrates individual engagement with the topic more deeply into the assignment.
You can copy this prompt directly:
Du bist eine didaktische KI, die Lernen durch eigene Beispiele fördert. Ich gebe dir eine bestehende Schulaufgabe. Erstelle eine neue Aufgabe, bei der:
- das gleiche Fachwissen vermittelt wird
- die Schüler:innen ein eigenes Beispiel aus ihrem Alltag,
ihrer Fantasie oder ihrem Umfeld entwickeln müssen
Baue folgende Elemente ein:
- eine kurze fachliche Erklärung (KI darf helfen)
- eine Aufgabe zur Übertragung: „Erfinde ein eigenes Beispiel, das zeigt, dass du das Thema verstanden hast.“
- eine Begründung: „Warum passt dein Beispiel zu diesem Thema?“
Achte darauf: Das eigene Beispiel muss so individuell sein, dass es nicht direkt von einer KI generiert werden kann.Here, too, you can simply add the original homework assignment afterward.
The advantage of this approach lies in its transfer efficiency. Learners should not only reproduce knowledge, but also apply it to their own context.
After all, it’s relatively easy to have an AI generate a response. Assessing whether an answer is appropriate, making your own connections, and applying specialized knowledge to a new situation requires significantly more independent effort.
Prompt 3: Check Understanding by Explaining
The third prompt is based on a well-known learning principle: We often don’t truly understand a topic until we can explain it clearly to someone else.
Here, too, students are allowed to use AI to begin exploring the topic. Afterward, however, they must explain what they have learned to a specific person on their own.
You can copy this prompt directly:
Du bist eine KI, die Hausaufgaben so gestaltet,
dass Schüler:innen Wissen erklären lernen.
Ich lade dir eine bestehende Aufgabe hoch.
Forme sie so um, dass die Schüler:innen:
- das Thema zuerst selbst verstehen (KI erlaubt)
- es anschließend für eine andere Person erklären müssen
Die neue Aufgabe soll enthalten:
- eine Zielperson (z. B. jüngeres Kind, neue Mitschülerin, Oma)
- klare Anforderungen:
- einfache Sprache
- eigenes Beispiel
- maximal X Sätze
- eine Reflexionsfrage:
„Was war schwierig beim Erklären und warum?“
Ziel ist zu prüfen,
ob das Thema wirklich verstanden wurde.Of course, the target audience can be tailored to the subject, age group, or specific task. Similarly, you can replace the “X” next to the maximum number of sentences with an appropriate number, or ask the AI to suggest a reasonable length on its own.
The actual learning achievement here lies in understanding, selecting, and restructuring information, and then conveying it to another person in a way they can understand.
AI Doesn’t Have to Be the Enemy When It Comes to Homework
These three prompts illustrate an important shift in perspective: We don’t have to focus exclusively on developing tasks that students must solve without AI.
Instead, we can design tasks that remain challenging even with AI.
Instead, we can place greater emphasis on skills that require active engagement: reflection, reasoning, application, individual decision-making, and the ability to explain knowledge clearly.
For teachers, this does not mean having to develop all teaching materials from scratch. Existing tasks can continue to serve as a starting point. With the right prompt, the AI can first generate suggestions for a new task. You can then review these and adapt them to your class, the learning objective, and the appropriate level.
These three approaches can also be combined
It gets especially interesting when you combine the three methods.
For example, an assignment might require students to first gather information using AI. Afterward, they must explain how they went about it and at what points they made decisions on their own. They then develop their own example and explain the topic to a specific target audience.
This gives rise to tasks in which AI can certainly serve as a tool—but it does not take over the entire thought process.
A prompt is no substitute for a pedagogical decision
As helpful as this approach is, the tasks generated by AI should not be adopted without review.
For example, check whether the new assignment actually aligns with the learning objective, is appropriate for the target age group, and requires meaningful independent work.
The requirement that something “cannot be generated directly by AI” should also be viewed in a nuanced way in practice. AI is now capable of generating a wide variety of examples, reflections, and explanations. It is therefore crucial to link tasks as closely as possible to the actual learning process, individual decisions, or concrete experiences of the learners.
AI gives you a starting point and can take a lot of the work out of redesigning existing tasks. However, the decision on how to teach it is up to you.
Conclusion: Not less AI, but better tasks
ChatGPT, Gemini, Kimi, and other AI systems aren’t going to just disappear from everyday learning. That’s why it’s worth thinking about how we can adapt our tasks to this new reality.
Instead of asking only how we can prevent students from using AI, we can ask a different question:
How can we design tasks so that learners can use AI while still having to think, understand, apply, and explain things for themselves?
The three prompts provide a practical starting point for this. You can copy it directly, add to your existing assignment, and see what new tasks result from it.
Have fun prompting and teaching!
Any further questions?
Do you have any questions? I am happy to support you, act as a sparring partner and answer your questions. I am always happy to receive your messages, preferably by WhatsApp message or e-mail.
