23 TrainingThings for AI-assisted instructional design

23 TrainingThings AI-assisted instructional design 1. Brainstorm course ideas e.g. by creating proand con lists for different topics. 2. Research about a topic / target group and find resources e.g. by using perplexity.ai or Googles Notebook LM. 3. Identify potential constraints & challenges. 9. Develop realistic scenarios, examples and problems targeted to your audience. 10. Check written content for typography, grammar & conciseness. 11. Get feedback (e.g. Did I miss something? Are the concepts introduced in logical order?). 12. Analyse the fit of course material to a goal and target group. AI will not replace you as an instructional designers —but using Large Language Models can help you make a course more relevant, speed up design, and enhance learning with active and social methods. Read page two for some recipes to create your own prompts. This guide presents our own best practices and does not assume to cover all possible aspects. Always evaluate AI outputs - and apply your own skills and creativity! 13. Break down problems into manageable sub-tasks with increasing complexity. 14. Create multiple-choice questions & good distinguishers (e.g. Create 8 answer options & mark the correct one with a *). 15. Enhance accessibility by creating image descriptions or transcripts. 16. Design gamified activities, scenario-based and role-playing activities tailored to course objectives. 17. Create images (e.g. DALL-E / Midjourney) to describe existing images in training material and create similar ones. 18. Create video scripts. 19. Create course announcements and participant communication. 20. Create targeted “What is in it for me” messages for the course based on the course material. 21. Produce reports based on learning analytics and feedback sheets. 22. Categorize text answers in feedback forms. 23. Create recommendations for the next round of course updates. Evaluation & Reporting 4. Generate learning objectives on different levels of Blooms Taxonomy. 5. Use AI to inform decisions, e.g. about appropriate course formats. 6. Generate outlines & agendas. 7. Brainstorm activity ideas for icebreakers, social learning, motivating participants, … 8. Choose the right format for quizzes, assessments and progress checks based on learning objectives. Implementation DOI: 10.5281/zenodo.17237958 Needs Analysis Course Design Course Development Media & Activity Creation 23 TrainingThings AI-assisted instructional design When creating your own prompts - there are some simple prompt frameworks you can use to optimize them: In addition, you can use rooms / projects or „custom GPTs“ to make the results more accurate, consistent and to not having to repeat instructions: •e.g. “give me percentage confidence scores for all outputs and create tables with pros and cons when providing different alternatives” In these specialized chats, you can upload course material and other documents to provide the model with more knowledge. In most tools, you can choose in your settings if your input will be used to further train the public model. Be careful to remove any personal data. Last but not least, Meta Prompts can give you AI-generated advice how to improve your prompts. Here are some examples: •“Suggest 3 better ways to phrase this prompt using the STAR method” •“My goal is to … How can you help me solve this task?” S - SITUATION (context / problem for background information) T - TASK (objective / goal to be achieved with the prompt) A - ACTION (concrete steps / how) R - RESULT (specific output requirements) C -CONTEXT (what you are trying to do & for whom) A -ASSIGNMENT (the task for the LLM) F -FORMAT (text, table, # of alternatives,..) E -EVALUATE (check sources, output, biases,..) Full Example Prompts: Create Learning Objectives “I want to create a course about data management planning for PhD students (no previous knowledge). The goal is to raise awareness of the benefits of DMPs for them and enable them to create their DMP with support from a Data Steward. Create 7 course objectives on different levels of Blooms Taxonomy.” Get feedback “Think like an instructional designer and give me feedback on the course material uploaded with regards to learner activity, exercises ..” Brainstorm activity ideas “For my DMP course for PhD candidates, I need some activity ideas for the introductory chapters and lower-level learning goals (with regards to Blooms Taxonomy). These objectives are…” Create scenarios and examples “Come up with five fictional scenarios of researcher requests for the deposition of their research data that are directed at research support staff. They should be from different disciplines.”