A new review article by Elisabeth Bauer, Samuel Greiff, Arthur C. Graesser, Katharina Scheiter and Michael Sailer, entitled “Looking Beyond the Hype: Understanding the Effects of AI on Learning”, has been published in the journal “Educational Psychology Review”. The article is available at the following link: https://doi.org/10.1007/s10648-025-10020-8

The publication offers a critical examination of current research and publication trends in the field of AI-supported learning. It identifies methodological weaknesses such as the frequent use of self-assessments or inadequate control groups. These can lead to distorted conclusions and overgeneralised statements about the benefits of AI in an educational context.

To systematically classify the potential impacts of AI on learning processes, the so-called ISAR model is presented (see Figure 1), which distinguishes between four types of impact: inversion, substitution, augmentation and redefinition. This creates a conceptual framework for a nuanced analysis of when and how AI supports cognitive learning processes – or, under certain circumstances, impairs them. The model is based on empirical findings and emphasises the importance of a well-thought-out didactic design, as well as individual and context-specific prerequisites for the successful use of AI in an educational context.

Figure 1: The ISAR model