Generative Artificial Intelligence in Professional Training of Teacher-Psychologists: Theoretical Analysis

Abstract

This article presents a theoretical analysis of the opportunities and limitations of using generative artificial intelligence (AI) in the professional training of teacher-psychologists in the Republic of Kazakhstan. The main contribution of the study is the introduction and theoretical grounding of the concept of a «zone of professional non-automatability» - a category of professional competencies (empathic contact, ethical judgment under uncertainty, interpretation of non-verbal behavior, intuitive recognition of crisis states) whose formation is fundamentally impossible through generative models and is achievable only through live interpersonal interaction. The proposed differentiation is theoretically grounded in M. Polanyi's concept of tacit knowledge and D. Schön's theory of the reflective practitioner. The applicability of the proposed framework is illustrated using the current curriculum of the educational program 6B01101 "Pedagogy and Psychology" at Abai Myrzakhmetov Kokshetau University, and is further supported by the authors' own empirical study (N=312), which demonstrated the effectiveness of generative AI primarily within the formalizable zone of competencies. The article also analyzes key directions for the use of generative AI in training teacher-psychologists (personalization of learning, modeling of professional situations, automation of routine tasks), the institutional context of digitalization of education in Kazakhstan, and potential risks of implementing the technology, including ethical aspects, academic integrity, and the language barrier, which underscores the importance of the national language model KazLLM. Conditions for the effective and safe integration of generative AI into the system of training teacher-psychologists are formulated, taking into account the national context of the Republic of Kazakhstan.