This paper presents a study of student perspectives on the didactic integration of digital learning elements into face-to-face university lectures and seminars. The overarching context is the use of the adaptive learning system (ALS) HASKI (short for Hochschullehre: Adaptiv, selbstgesteuert, KI-gestützt, i.e. Higher Education: Adaptive, Self-Directed, AI-Supported), which was tested in computer science teaching within a blended learning approach, more specifically in flipped classroom scenarios.

The joint project, consisting of three Bavarian universities, focuses on the exploratory integration of an AI-based ALS into higher education. The system was applied in a course for mechatronics students (2nd semester) and provided a variety of adaptive learning elements. HASKI generates individual learning paths with AI based on learning behavior and learning styles. It is designed to promote personalized, self-directed learning as support for blended learning scenarios.

The method chosen for data collection is a qualitative content analysis based on exploratory interviews with a semi-structured set of questions. Four central analysis criteria were considered: acceptance and perception of the learning elements, integration into lectures and exercise sessions, didactic coherence, and suggestions for improvement.

The results show that students prefer explanatory scripts, interactive tasks, and quiz elements that require a certain degree of reflection. In addition, the respondents are largely convinced of the integration into the seminar, as the HASKI system adequately guides them for what to prepare for. When it comes to embedding the system into lectures, especially in the form of question rotation in small groups, the feedback is ambivalent. Although the potential for in-depth learning was recognized, the low participation of fellow students was critically reflected upon.

A particular difficulty mentioned was the lack of coordination between self-study and classroom attendance. While some participants saw this as a discrepancy of coherence, others viewed the adaptive system as a balancing factor. There was a desire for a clearer time structure and more in-depth materials that go beyond mere repetition.

Overall, the findings provide initial indicators of successful aspects in the integration of adaptive systems into classroom teaching. The continuous development of a clear didactic division of roles between self-study and classroom phases is of central importance here. Further research with larger samples, if necessary, could provide more comprehensive insights.

Keywords: Adaptive Learning System, Blended Learning, Flipped Classroom, Higher Education, Didactics, Learning Path Personalization, Student Perspective, Qualitative Content Analysis, e-learning, Digital Learning.