Recent and ongoing research in the fields of Educational Psychology and Empirical Educational Research will be presented and discussed in the research colloquium. In addition to scientists from the University of Augsburg, researchers from other research institutions will also report on their work. All colleagues, doctoral researchers and students are cordially invited.

Artificial intelligence (AI) can shape learning in both beneficial and undesired ways. For example, according to the ISAR model, AI-supported learning may result in inversion effects when learners offload essen tial cognitive activities to AI and engage more superficially. However, AI-enhancement may also lead to positive effects, such as augmentation effects, when AI provides additional support that helps learners engage more effectively with the task and its content. Which type of effect emerges is likely not only a matter of the tool and instructional context, but also of how learners use AI during the learning task. For this reason, examining learning processes is essential. At the same time, these processes may depend on learners’ cognitive prerequisites, such as prior knowledge, which can shape how productively AI is used in the learning task. This presentation provides an overview of a series of studies on learning in the context of reasoning about socio-scientific issues with and without AI support. The studies address complementary questions, including how AI-supported reasoning relates to cognitive load and reasoning outcomes, how learners interact with AI during the learning process as reflected in their chat data, and how differences between learner groups may point to the role of cognitive prerequisites in AI-enhanced learning. Together, the presented results contribute to a conceptual and empirical understanding of when AI support may hinder or foster learning.