When Does AI Help Learning - and When Does It Replace It?

Generative AI in education requires a shift from binary "good or bad" debates toward understanding the specific conditions that foster learning. The report "Young People, Learning and Generative AI" identifies five interacting factors—the learner, the tool, the task, social configurations, and institutional context—that determine whether AI enhances or hinders educational outcomes. While AI can improve engagement and reduce anxiety, it introduces new metacognitive demands, requiring students to discern when to delegate tasks. Strategic offloading, where AI handles extraneous loads to focus on core learning objectives, offers a productive path forward. Teachers remain essential in this ecosystem, acting as facilitators who interpret student needs and design tasks that prevent detrimental shortcuts. Future research must prioritize localized, primary-school-level studies to move beyond the current concentration of data in higher education and international contexts.