Keyword: «automatic speech recognition (asr)»

The relevance of the study is determined by the rapid integration of digital educational technologies into foreign language teaching against the background of insufficient research into their impact on the cognitive mechanisms of spontaneous foreign language speech (SFLS) production. The pace of EdTech product and Automatic Speech Recognition (ASR) system dissemination outstrips scientific analysis, generating a methodological contradiction: marketing metrics supplant psycholinguistic analysis of actual changes in the learner's speech production architecture. The purpose of the article is a systemic analysis of the mechanisms by which digital tools affect the learner's cognitive functions in developing SFLS, and the development of a theoretically grounded integration model for distributing functions between the digital environment and the teacher. The leading approach is psycholinguistic analysis within A. Baddeley's multicomponent working memory model and W. Levelt's speech production model; the activity approach, comparative, typological, and systematising methods, and modelling are also employed. Main results: (1) three key concepts clarified – SFLS, cognitive functions, digital learning tools; (2) a symmetrical systematisation of eight advantages and eight limitations of digital tools correlating each position with a specific cognitive mechanism; (3) a typology of six ASR system failure types with their negative cognitive effects characterised; (4) the irreplaceability of live pedagogical interaction justified through three functions unrealisable in a digital environment; (5) a four-stage integration model proposed (orientation-diagnostic, training, productive, reflective stages), distributing functions between digital tools and the teacher on the principle of cognitive expediency. Theoretical significance: systematisation of psycholinguistic mechanisms by which digital technologies affect the learner's cognitive functions, with an explanatory framework correlating EdTech capabilities with speech production operations. Practical significance: methodological recommendations for the critical application of ASR tools and for the organisation of a hybrid educational environment that ensures an optimal balance between automated and live speech practice.