From Prompt to Practice: Design AI-Supported Tasks

Reading time: approx. 5 min

Congratulations! You have now gone through the most important parts of the course "How to Think Like an AI". You have learned about AI's basic principles, tokens, context window, different prompting techniques, and how you handle hallucinations. In this final moment, we gather all these techniques and principles so that you can design and implement effective, AI-supported lesson activities in your own teaching. The goal is for you to feel confident integrating AI as a tool to enhance both your and your students' learning.

1. Choose your lesson idea

Start by identifying a concrete lesson task or a teaching moment where you see potential for AI support. Think about where AI can free up time, offer new perspectives, or individualize learning. Some examples can be:

• Creating quiz questions: Quickly generate multiple-choice questions, open questions, or true/false statements based on a specific subject area.
• Formative feedback on student texts: Using AI as a first step to give students general feedback on text structure, grammar, or idea development.
• Summary of read material: Letting AI help students condense long texts into shorter summaries, focus on key concepts, or extract important information.
• Role-playing tasks: Using AI as a conversation partner for students to practice language skills, debate different viewpoints, or explore historical perspectives.
• Differentiate materials: Adapt the same content to different reading levels or linguistic abilities in the class.

2. Apply your knowledge from the course

When you design your AI-supported task, actively use what you have learned:

Tokens and context window (Moments 2 & 3)

• Keep prompts focused: Avoid overloading the AI with too much information at once.
• Prioritize important information: Place the most critical instructions early in the prompt.
• Break down complex tasks: If the task is too large, break it down into smaller steps.

Zero-shot and few-shot prompting (Moments 4 & 5)

• Start simple: First test zero-shot to see if the AI understands the task directly.
• Add examples when needed: If the result is not good enough, use few-shot with 1-3 concrete examples.
• Adapt to student level: Make sure examples and language match your students' abilities.

Temperature and creativity (Moment 7)

• Low temperature for facts: When precision is important (for example, grammar checking, factual questions).
• High temperature for creativity: When you want the AI to generate varied, creative answers (for example, story ideas, brainstorming).

Handle hallucinations (Moment 8)

• Double-check facts: Always verify AI-generated factual information.
• Teach students source criticism: Make it part of the exercise to verify the AI's answers.
• Use RAG techniques: When possible, give the AI specific texts to base answers on.

3. Practical design template for AI-supported tasks

Use this step-by-step template when you create your own tasks:

Step 1: Define goals and purpose

• Pedagogical goal: What should students learn?
• AI's role: How should the AI support this learning?
• Students' role: What are students expected to do independently?

Step 2: Design the prompt(s)

• Clear instruction: Describe clearly what the AI should do.
• Context information: Give relevant background (grade, subject, specific requirements).
• Format and structure: Specify how the answer should be structured.
• Constraints: Set clear limits for length, style, or content.

Step 3: Test and iterate

• Run the prompt: Test it with the AI before you use it with students.
• Evaluate the result: Does it meet your expectations?
• Adjust: Fine-tune prompt, temperature, or approach based on the results.

Step 4: Plan for safety and ethics

• Fact-checking routine: How should students verify the AI's answers?
• Source criticism moment: Integrate discussion about AI's limitations.
• Originality: Make sure AI is used as a tool, not to replace the student's own reflection.

4. Three concrete examples of AI-supported tasks

Example 1: English teaching - Text analysis with AI support

Pedagogical goal: Students should be able to analyze and interpret a short story.

AI prompt template:

You are a pedagogical AI assistant for English teaching at high school level.

Read the following text: [STUDENT'S CHOSEN TEXT]

Help the student prepare their text analysis by:
1. Identifying 3 important themes in the text
2. Pointing out 2 stylistic devices the author uses
3. Suggesting 3 discussion questions that the student can explore further

Write in English and adapt the language for high school students. Do not give ready answers but help the student come up with their own thoughts.

Student task: Students use the AI as a discussion partner to prepare their own text analysis, but write the final analysis themselves.

Example 2: Mathematics - Problem-solving with step-by-step guidance

Pedagogical goal: Students should develop problem-solving strategies in algebra.

AI prompt template:

You are helping a student in grade 9 with algebraic problems.

Problem: [STUDENT'S SPECIFIC PROBLEM]

Instead of giving the direct solution:
1. Ask a leading question that helps the student get started
2. If the student answers, confirm and ask what the next step could be
3. Give hints if the student gets stuck, but let the student do the calculations

Use encouraging language and adapt to English school mathematics.

Student task: Students "converse" with the AI to work through the problem step by step, which trains both mathematical understanding and metacognition.

Example 3: Social studies - Perspective exercise with role-playing

Pedagogical goal: Students should understand different actors' perspectives on a social issue.

AI prompt template:

You should act as [SPECIFIC ROLE: for example, "an environmental activist", "a business leader", "a politician"] in a discussion about [CURRENT SOCIAL ISSUE].

Base your answers on this role's likely opinions and interests. Use arguments that this person would use, but avoid extreme or offensive positions.

Answer in English and maintain a respectful, factual tone even when you argue for your position.

Student task: Students "interview" different AI roles to gather perspectives, which they then should analyze and compare in their own reflection.

5. Quality control and evaluation

When you have implemented your AI-supported task, evaluate it according to the following criteria:

Pedagogical effectiveness

• Did students' understanding or performance improve?
• Did you save time that could be used for more meaningful activities?
• Did students become more engaged in the subject?

Technical function

• Did the AI give consistently useful answers?
• Did prompts need to be adjusted during the lesson?
• Did students have technical problems?

Ethical aspects

• Did students learn to critically examine the AI's answers?
• Did students retain ownership of their own learning?
• Were the AI's limitations discussed openly?

6. Next steps in your AI journey

Now that you have completed the course and started integrating AI in your teaching:

Continue experimenting

• Test new AI tools and functions that become available
• Adapt techniques to different subjects and situations
• Collaborate with colleagues to exchange experiences

Deepen knowledge

• Learn more about specialized AI tools for education
• Explore advanced techniques like RAG (Retrieval Augmented Generation)
• Participate in professional development in AI and pedagogy

Contribute to the staff

• Share successful tasks and prompt templates
• Discuss challenges and solutions
• Help other teachers get started with AI

Reflection exercise: Your first AI-supported lesson

To consolidate what you have learned, now plan a concrete AI-supported lesson activity:

1. Choose subject and goal: What should students learn?
2. Design prompt: Write a prompt according to the principles from the course
3. Test: Run the prompt and evaluate the result
4. Plan implementation: How will you introduce this to students?
5. Prepare safety: What control mechanisms do you need?

Successfully completed course! You now have basic knowledge of how AI models work and practical tools to integrate them in your teaching in a responsible and effective way. AI is not a substitute for good pedagogy, but a tool that can enhance and improve what you already do well. Use it wisely, and remember that your pedagogical knowledge and judgment are always the most important.