Abstract:

The increasing integration of genAI in higher education challenges both students and educators to develop not only technical proficiency but also resilience: The ability to engage critically, reflectively, and autonomously with AI tools. This paper presents the design and empirical evaluation of the course "License for Digital Competence", a cross-disciplinary, problem-based learning format that aims to foster digital competences [DigComp 2.2] and focuses on the responsible use of AI among students.

The semester-long, 5 ECTS course guides students through six problem-based phases in small groups. Each phase presents a case scenario that addresses real-world challenges of digitalization. Students autonomously choose tools [AI or other] to explore, apply, and reflect on their learning process. The pedagogical design emphasizes self-organization, peer collaboration, formative feedback, and continuous reflection through:
1) Consultation sessions with the educators. The instructor acts as a learning facilitator rather than a content authority, creating space for experimentation, critical dialogue, and collaborative meaning-making.
2) Reflective videos that go along with their artifacts. These are guides by leading questions.
3) Discussion within their groups while creating the solution for the problem-based case study phases.
4) Feedback for each phase by the educator. This means that groups receive feedback once every 2 weeks.

The concept of AI resilience is operationalized as students’ ability to:
a) make conscious and situational decisions about when and how to use AI tools
b) assess the quality and reliability of AI-generated outputs
c) maintain a sense of self-determination and agency in technology-mediated learning environments
d) take responsibility for their artifacts by claiming them in an AI-transparancy statement

To evaluate the course’s impact, all current and former cohorts (approximately n=140 over 5 semesters) were surveyed using a quantitative questionnaire, with descriptive analysis employed. Students reported increased awareness of AI’s potentials and limitations, greater confidence in making informed decisions about tool use, and enhanced capacity to critically reflect on AI outputs. Consultations, reflection videos, and group discussions were identified as the most influential elements supporting this development.

Findings indicate that AI resilience emerges not from prohibiting or uncritically embracing AI, but through structured reflection, social learning, and process-oriented assessment. Evaluating how students learn, rather than what they produce, proved essential. The results suggest that higher education should shift from summative product evaluation toward formative, dialogue-based assessment models that value reflection, process documentation, and ethical awareness.

The paper concludes with practical recommendations for educators: foster open exploration, model reflective AI use, and integrate structured reflection cycles into problem-based designs. By embedding AI reflection within authentic learning contexts, universities can cultivate digitally competent, self-determined graduates capable of navigating AI-rich learning and working environments with critical awareness and resilience.

Keywords:

AI Resilience, Didactical Design, Reflective Learning, Higher Education, Digital Competence.