Many existing digital tutoring solutions rely on standardized learning paths and automated AI interactions, with limited understanding of how personalized content and didactic strategies actually affect learning. EduMind follows a research-driven approach to investigate how real-world, interest-based learning content and different didactic explanation styles influence motivation, comprehension, and learning outcomes. Based on these insights, human-centered AI mechanisms are developed that actively involve teachers and meaningfully extend existing digital learning platforms.

EduMind aims to develop an AI-based tutoring module that seamlessly integrates into existing digital learning platforms (e.g., Moodle). The module supports learners through personalized, real-world explanations tailored to individual interests and combines them with teachers’ didactic expertise. This approach is intended to sustainably increase motivation, comprehension, and learning success.

EduMind combines state-of-the-art AI technologies with pedagogical concepts:

• Large Language Models (LLMs) for generating personalized learning content
• Retrieval Augmented Generation (RAG) for context-aware answers to learning questions
• OCR and semantic analysis to process existing teaching materials
• Prompt engineering using teachers’ didactic examples
• Text-to-image models to visualize complex topics
• Automatically generated assessments to identify learning gaps