Blended learning, conceptualized as the integration of online and face-to-face instructional modalities with student-centered pedagogical elements, provides unprecedented flexibility and accessibility within contemporary educational environments. However, in bilingual contexts its pedagogical potential is often constrained by a range of structural challenges associated with the unequal representation of languages in digital environments. Such imbalances frequently contribute to academic inequality and increase the cognitive load experienced by learners. Empirical evidence derived from the implementation of blended learning in technical higher education institutions [3] demonstrates the substantial capacity of this approach to foster the development of digital competencies and technologically mediated learning practices.

In this regard, artificial intelligence (AI) technologies offer new opportunities for addressing these challenges, functioning as a catalyst for the development of genuinely adaptive and personalized educational ecosystems. Within the context of teaching technical disciplines, such as computer science, where terminological precision is of critical importance, intelligent systems can facilitate digital linguistic parity. Through this process, bilingualism can be transformed from a source of additional cognitive burden into a significant pedagogical resource that enhances learning effectiveness and conceptual comprehension.

The term “blended learning” is widely employed in both educational practice and scholarly discourse. Although the concept has been defined and redefined by numerous researchers, none of the existing interpretations fully captures the structural complexity of the blended learning model or adequately explains how its various components interact to produce a coherent and integrated educational outcome. Within the framework of the present study, and following the interpretation proposed by Graham et al., blended learning is conceptualized as a pedagogical model that integrates two principal instructional approaches: a traditional format grounded in direct, face-to-face interaction between students and instructors, and a computer-mediated mode of instruction that utilizes digital technologies to facilitate and support the educational process.

The empirical basis of the study consisted of scholarly publications produced by both international and domestic researchers, focusing on the theoretical foundations and conceptual development of blended learning models. As the primary research method, meta-analysis was employed, allowing for the systematic synthesis and integration of findings from multiple independent studies addressing the same research problem. The meta-analytical procedure comprised several stages: the formulation of the research objective; the systematic search and selection of relevant studies that met predetermined inclusion criteria; the extraction and compilation of data from the selected sources; the application of statistical techniques for data analysis; and, finally, the interpretation of the results with an emphasis on assessing their significance and implications for the field.

A review of the scholarly literature makes it possible to identify several key research directions within which the issue of integrating artificial intelligence into bilingual and blended learning is examined. This problem area lies at the intersection of pedagogy, linguistics, computer science, and educational policy, which determines the interdisciplinary nature of the present study.

In the context of the ongoing digital transformation of education, particular attention has been given to the theoretical foundations of bilingual education. The works of prominent international scholars, such as Jim Cummins, emphasize the need to reconsider pedagogical practices under conditions of rapid digitalization. Cummins [7] highlights the potential risks associated with the expansion of digital inequality, particularly for minority languages, and argues for the importance of developing inclusive educational environments in which technology functions not as a substitute for linguistic and cultural identity, but rather as a mechanism that strengthens and supports it.

The theoretical framework of translanguaging has been further elaborated in the works of Ofelia García. Within this perspective, bilingual language practices are not conceptualized as two separate linguistic systems but as a unified and flexible communicative repertoire. Research in the field of adaptive learning systems conducted by scholars such as Tzung‑Shiung Yang, Gwo‑Jen Hwang, and Stephen J.H.Yang [3] demonstrates the significant potential of digital technologies for the personalization of the educational process through the consideration of multiple learner characteristics. Nevertheless, most existing technological solutions are designed for standard language pairs and rarely account for the linguistic particularities of agglutinative languages characterized by rich morphological structures. A substantial contribution to the investigation of regional linguistic and educational specificities has been made by Russian scholars. In particular, the research conducted by D.Sh.Suleymanov, R.A.Gilmullin, and A.R.Gatiatullin provides a detailed analysis of contemporary linguistic resources and information technology developments related to the Tatar language. Their work identifies a critical imbalance in the availability of digital educational resources and outlines potential strategies for developing linguistic corpora to support automated text processing. Furthermore, studies devoted to the organizational and didactic typology of blended learning models conducted by V.I.Blinov, E.Yu.Yesenina, and I.S.Sergeev examine the didactic potential of digital technologies for adaptive learning. These scholars identify several promising directions, including the intelligent adaptation of educational content, diagnostic assessment of learning outcomes, and the development of methodological support systems for educators operating in digital learning environments. With regard to blended learning specifically, the methodological aspects of its implementation in higher education institutions, including rotation models and the organization of online interaction, have been examined in the studies of T.Yu.Pletyago, A.S.Ostapenko, and S.N.Antonov [1]. Their work contributes to the development of practical frameworks for integrating blended learning strategies into university-level educational practice.

The initial conceptualization of blended learning as a simple mechanical combination of traditional and distance education has gradually evolved into a more sophisticated theoretical interpretation. Contemporary perspectives emphasize its multidimensional nature and pedagogical complexity. In the Glossary of Digital Didactics Terms and Concepts, blended learning is defined as “the integration of two or more distinct modes, forms, and methods of instruction—traditional and electronic, face-to-face and distance, synchronous and asynchronous, formal and informal within a unified educational process” [6]. This definition reflects the systemic integration of diverse instructional modalities within a single pedagogical framework.

A significant contribution to the development of the blended learning paradigm was made by scholars such as Curtis J.Bonk, Charles R.Graham, Jay Cross, and Michael G.Moore. In their collective work The Handbook of Blended Learning: Global Perspectives, Local Designs, the authors articulated one of the earliest comprehensive definitions of the concept. According to their interpretation, blended learning represents an instructional system grounded in the integration of face-to-face (in-person) instruction and computer-mediated learning environments. This formulation highlights the complementary relationship between conventional classroom interaction and technology-supported educational practices.

Further theoretical refinement of the concept was proposed by Heather Staker and Michael B. Horn in their study Classifying K–12 Blended Learning. The authors conducted a critical examination of existing definitions of blended learning and suggested their revision and expansion in response to contemporary educational innovations [5]. Their methodological approach was guided by two fundamental principles: (1) the development of flexible conceptual definitions capable of maintaining relevance amid continuous technological advancement, and (2) the deregulation of rigid normative constraints that may hinder pedagogical innovation [1]. An important outcome of their research was the creation of a detailed classification of blended learning models, which provides a structured framework for analyzing and implementing various forms of technology-enhanced instruction.

The growing relevance of blended learning within modern educational paradigms is also reflected in the works of prominent scholars in educational research. For instance, L.L.Salekhova and her colleagues present a comprehensive synthesis of theoretical perspectives and empirical findings that demonstrates the effectiveness of a systemic approach to the design of learning environments. Their study employs the CABLS (Comprehensive Analysis of Blended Learning Studies) methodology, which enables a multi-level analytical examination of blended learning practices. Particularly valuable is the identification of key determinants of successful blended learning implementation, including technological infrastructure, methodological support, and psychological-pedagogical facilitation [2].

Empirical evidence obtained in these studies indicates a significant improvement in academic performance when online and offline instructional formats are effectively integrated. Such findings are especially relevant within bilingual educational contexts, where instructional flexibility and adaptive pedagogical strategies are essential for optimizing the learning process and supporting diverse linguistic and cognitive needs of learners.

In the field of international research focusing on artificial intelligence technologies and adaptive learning, significant scholarly attention has been directed toward the effectiveness of Intelligent Tutoring Systems (ITS)—computer-based instructional systems designed to emulate the functions of a human tutor through the provision of personalized instruction and adaptive feedback. Within this research domain, K. VanLehn conducted a comparative analysis of various tutoring systems and concluded that ITS capable of adapting not only to learners’ subject-matter knowledge but also to their individual cognitive characteristics demonstrate the highest levels of instructional effectiveness [4]. These findings substantiate the argument that artificial intelligence can function as an effective support tool for bilingual learners and confirm that well-designed adaptive learning systems are capable of producing meaningful educational outcomes.

Robert Godwin-Jones, in his examination of the evolution of educational technologies, anticipates a shift from static instructional programs toward hybrid intelligent learning environments, in which artificial intelligence operates as a personal learning assistant. Within such environments, AI provides contextualized guidance and facilitates the construction of individualized learning ecosystems tailored to the needs of each learner. Foundational theoretical contributions by J. Sweller on Cognitive Load Theory and by R. Baker on Educational Data Mining (EDM) [1] constitute the methodological framework for the development of adaptive systems aimed at minimizing excessive cognitive load among bilingual students. Educational Data Mining applies data mining techniques to analyze large volumes of learner-generated educational data. Through these analytical processes, EDM enables researchers and educators to identify the most effective learning strategies and design personalized learning pathways for individual students.

Importantly, the principles of cognitive load management are directly applicable to the design of instructional interfaces and educational materials. Such designs must reduce cross-linguistic interference and facilitate the comprehension of complex technical concepts, particularly for bilingual learners who simultaneously process information across multiple linguistic systems.

Another significant direction in the advancement of educational artificial intelligence involves the development of adaptive learning systems. Research conducted by T.-S. Yang, Gwo-Jen Hwang, and Steven J. H. Yang [3] demonstrates that the highest levels of instructional effectiveness are achieved when adaptive systems consider a multidimensional set of learner characteristics, rather than relying on a single isolated parameter. Within the context of the present study, this finding suggests that the proposed model should implement adaptive mechanisms not only on the basis of subject knowledge but also in relation to the learner’s individual linguistic profile. Such a multifactorial adaptive approach enables the construction of a genuinely personalized educational trajectory. Within this framework, artificial intelligence dynamically selects linguistic scaffolding, content formats, and instructional sequencing. As a result, cognitive load is reduced, cross-linguistic interference is mitigated, and the overall effectiveness of learning complex academic material is significantly enhanced.

Thus, methodological approaches applied in foreign language instruction include the systemic, activity-based, learner-centered, and competence-based approaches. These approaches determine the overall orientation for designing a structural–functional model of the educational process, aimed at identifying the systemic relationships among the key components of school-based language learning. Furthermore, they promote students’ active engagement in the processes of cognition and subject-matter acquisition, while simultaneously fostering the holistic development of the learner’s personality and the enhancement of individual abilities in accordance with the expectations and demands of the surrounding social environment. The structural–functional model of the blended learning process comprises several interrelated components: the target, content, technological, and assessment-resultative blocks. These components function in a relationship of mutual interdependence and are aligned with the social demand articulated by parents, legal guardians, and the broader community. The identified levels for evaluating the effectiveness of this structural–functional model provide a framework for assessing the successful implementation of blended learning within the distance educational environment of schools.