Beyond Outcomes: How AI Tools Reshape the Multilingual Learning Process Through Cognitive, Affective, and Autonomy-Building Mechanisms

Abstract

The rapid integration of artificial intelligence (AI) into language education has attracted growing scholarly attention. However, most of the existing studies have focused on measurable learning outcomes, while the specific mechanisms of AI in reshaping the learning process have not been thoroughly explored or substantiated. Especially in multilingual learning environments, learners typically need to master multiple language systems and face multiple cognitive and emotional challenges. According to the existing systematic literature review, this paper proposes a three-threshold framework (including cognitive threshold, emotional threshold, and autonomy threshold). The analysis shows that the greatest contribution of AI to multilingual learning lies in re-establishing conditions for continuous learning, rather than directly influencing the attainment of learning outcomes. Implications for multilingual pedagogy and AI tool design are also considered, along with the limitations of this non-empirical approach.