Text as a Tool. The Effects of Using Image-Generating AI in German University Contexts
Text as a Tool. The Effects of Using Image-Generating AI in German University Contexts

Summary/Abstract: This article investigates how German university students learn to use image-generating artificial intelligence and its implications for educational practice. Using semi-structured interviews and think-aloud protocols, twelve students (aged 18-26) from media and social science programs at three German universities created personal and academic images using Midjourney. Through descriptive thematic analysis, findings reveal that image-generating AI creates initial excitement and motivation while fostering soft skills including persistence, precise communication, and openness to new experiences. However, some students experienced frustration when attempting to realize specific visual concepts and showed susceptibility to distraction. Notably, students with broader general knowledge in art history and technical terminology achieved more successful outcomes. Further, the study identifies a fundamental shift in university learning paradigms: whereas academic work traditionally progresses from images to textual analysis, image-generating AI reverses this process, positioning text as a tool for visual creation. This has implications for curriculum designs, suggesting the need for enhanced visual literacy, expanded general education requirements, and explicit instruction in AI bias recognition. For educational practice, image-generating AI can serve as a valuable pedagogical tool for increasing engagement while requiring careful scaffolding to prevent distraction and ensure critical evaluation of outputs. The study contributes foundational insights for developing pedagogical frameworks that harness the potential of image-generating technologies in education.

• Details
• Contents