The rapid integration of artificial intelligence (AI) into language education has fundamentally transformed how learners acquire, process, and use second languages. From intelligent tutoring systems (ITS) that provide personalized feedback in real time, to natural language processing (NLP) tools such as ChatGPT and Grammarly that scaffold writing and speaking, AI technologies are reshaping the sociocultural landscape of language learning. These tools are not merely add-ons to traditional pedagogy; they create new mediating artifacts within the learning environment, altering the dynamics of social interaction, scaffolded assistance, and collaborative meaning-making. Despite the proliferation of AI-enhanced language learning tools, the theoretical understanding of how these technologies mediate language learning outcomes remains fragmented. Most existing studies adopt a techno-centric perspective, focusing on feature evaluation or effectiveness comparisons while neglecting the deeper sociocultural mechanisms through which AI tools facilitate or hinder learning. This gap is particularly pronounced when considering the mediating roles of sociocultural factors such as social interaction within AI environments, the zone of proximal development (ZPD) created by adaptive scaffolding, and the co-construction of knowledge through human-AI collaboration. Sociocultural Theory (SCT), originally formulated by Vygotsky and subsequently expanded by scholars such as Cole, Wertsch, and Lantolf, provides a robust theoretical lens for examining how learning occurs through socially mediated activity. SCT posits that cognitive development is fundamentally embedded in social interaction, cultural tools, and historically situated practices. Applying SCT to the context of AI-enhanced language education offers a compelling framework for understanding how AI tools function as cultural artifacts that mediate language learning within social and collaborative contexts.
Huong et al. (Thu,) studied this question.