Beyond the Educational Loop: Generative Struggle in AI-Mediated English Language Acquisition, Além do Ciclo Educacional da IA: Esforço Generativo na Aprendizagem da Língua Inglesa Mediada por Inteligência Artificial

<jats:p>Background: This study develops generative struggle as a pedagogical framework for AI-mediated English language acquisition, addressing the cognitive risks associated with AI-driven task completion and cognitive offloading. Objective: The study aims to: (1) identify psychological barriers that may influence learners’ engagement with their own cognitive resources during AI-mediated writing tasks, and (2) conceptualise the teacher’s role as an epistemic mentor who promotes generative struggle, divergent thinking, learner agency, and meaningful language development in GenAI-assisted writing environments. Methods: The study adopts a conceptual and theoretical synthesis approach, integrating perspectives from cognitive psychology, second language acquisition, English language teaching, and educational technology. The inquiry draws on five established traditions: the generation effect, desirable difficulties, writing-to-learn, error-based learning, and cognitive offloading. These perspectives are considered alongside the learner-centred vision of NEP 2020 to examine the relationship between cognitive effort and meaningful language development. Results: The analysis suggests that excessive reliance on algorithmic efficiency may erode cognitive engagement by bypassing the generative struggle central to language acquisition. The proposed framework therefore prioritises generative struggle and guided self-correction over mere task completion, repositioning the teacher as an epistemic mentor who fosters reflective, metacognitive, and critical engagement beyond the “impressiveness trap.” Rather than rejecting automation, the framework advocates pedagogically mediated AI integration that supports generative learning while limiting cognitive outsourcing. These findings provide a basis for preliminary instructional directions adaptable to learners’ proficiency levels, institutional contexts, and assessment designs. Conclusion: The paper conceptualizes generative struggle as an integrative construct that synthesizes five established cognitive traditions to explain learning under AI-mediated conditions. The concepts of the AI education loop and the impressiveness trap further illuminate emerging tensions between AI-assisted performance, cognitive engagement, and genuine learning. Together, these concepts provide a conceptual basis for understanding how AI-mediated educational environments can support meaningful learning without displacing the cognitive effort essential to language development.</jats:p>