INTEGRATING ARTIFICIAL INTELLIGENCE INTO FOREIGN LANGUAGE INSTRUCTION: CAPABILITIES AND DIDACTIC CONSTRAINTS

Abstract

This study explores the prospects of employing generative neural networks (ChatGPT, Gemini) alongside speech synthesis technologies within the context of second language acquisition. It identifies three primary applications of AI: as a simulated dialogue partner, an individually tailored instructional tool, and a support mechanism for educators. The article draws on meta-analytical findings (Huang et al., 2022; Kohnke, 2023), which indicate notable reductions in oral communication apprehension (Cohen’s d up to 1.06) and enhanced speaking fluency following regular interaction with AI-based systems. Consideration is also given to the phenomenon of "illusory competence" and suggested instructional strategies to counteract it, such as training learners in metacognitive engagement with neural network models.