Veranstaltungsprogramm

The last decade has seen a surge of interest in Virtual Exchange (VE) and Collaborative Online International Learning (COIL) as mechanisms for Internationalization at Home (IaH) of Higher Education, bringing intercultural collaboration into mainstream curricula without the logistical and financial barriers associated with physical mobility. Across disciplines, VE/COIL supports interactional authenticity, mutual perspective‑taking, and cross‑institutional teamwork. In addition to cognitive gains, participants report key socio‑emotional developments for global citizenship such as heightened curiosity, empathy, and tolerance for ambiguity.

Assessment of VE/COIL projects is difficult because outcomes are multidimensional, contextual, and dynamic. Intercultural competence comprises attitudes, knowledge, skills, and internal/external outcomes. Evidence accrues in interaction, through written and spoken discourse, collaborative artifacts, and reflection. Traditional surveys and activity counts underrepresent qualitative growth; they are ill‑suited to capturing behavioral transfer over time or linking participation to organizational results. Methodological pluralism is essential, yet reflective and qualitative approaches are labor‑intensive and challenging to scale across large cohorts and multi‑institutional contexts.

The Kirkpatrick Model offers a well‑established scaffold to organize questions, methods, and evidence from immediate experience through long‑term impact. On the other hand, Artificial Intelligence (AI) can provide continuous feedback loops, reduce manual coding burdens, and uncover patterns in large datasets, while human‑in‑the‑loop interpretation preserves pedagogical judgment and contextual nuance.

This paper suggests an evaluation framework that augments each level of the Kirkpatrick Model (Reaction, Learning, Behavior, Results) with AI methods to support VE/COIL evaluation and assessment by enhancing sensitivity, timeliness, scalability, and data‑driven insights. At Level 1 (Reaction), natural language processing (NLP) and adaptive surveys capture participant sentiment and perceived relevance in real time. Level 2 (Learning) integrates adaptive assessments, simulations, and learning analytics to track knowledge and skill acquisition against intended objectives, including intercultural competence. Level 3 (Behavior) leverages behavioral analytics and reflection analysis to identify application and transfer of skills across authentic tasks and over time. Level 4 (Results) connects VE/COIL participation to organizational goals via predictive modeling and data mining, surfacing impacts on performance, collaboration, retention, and cultural understanding. The framework offers institutions a practical pathway to nuanced, scalable, ethically governed evaluation and outlines priorities for mixed‑methods validation, equity‑minded adoption, and institutional sustainability.