META-GEN: Metacognitive Engagement in Textual Activities with Generative AI

Abstract

Generative Artificial Intelligence (GAI) is profoundly reshaping learning and text production practices, offering new forms of cognitive support while raising questions about its impact on metacognitive processes and learners’ reflective autonomy. The META-GEN project explored the relationship between GAI use and the activation of planning, monitoring, and evaluation strategies in comprehension and writing tasks. A quasi-experimental design with a control group involved 296 adult learners engaged in argumentative text production, with or without GAI assistance. Data were collected through the Metacognitive Awareness Inventory (MAI), a digital competence questionnaire, an ad hoc survey on metacognitive strategy use, and a post-task comprehension test. Descriptive and inferential analyses revealed high levels of metacognitive awareness and solid monitoring and evaluation skills, but weaker strategic planning in GAI-mediated tasks, suggesting a selective cognitive delegation effect. No significant differences emerged in comprehension between groups, indicating that GAI does not reduce learning quality but modifies its underlying processes. GAI appears as a conditional metacognitive facilitator: it can foster self-regulation and reflection when intentionally integrated into an educational design that makes control strategies explicit. The study highlights the need to promote AI literacy and metacognitive scaffolding to balance technological efficiency with reflective responsibility in learning processes.