Prospects for integrating artificial intelligence into the system of pedagogical mentoring: Results of a regional study

Abstract

The relevance of this study stems from the need to integrate artificial intelligence (AI) into pedagogical mentoring, as outlined in the National AI Development Strategy until 2030. Despite the deep penetration of generative AI in education, there is no comprehensive understanding of the social and institutional factors shaping teachers' readiness to use AI in mentoring. This study aims to analyze the readiness of educators and educational institutions for such integration. The methodology draws on the SAMR model, SCOT, JTBD, and the AI-TPACK concept. The empirical study (2024-2025) was conducted in four educational organizations in the Tomsk region, varying in profile and level of digitalization. The sample included 371 respondents: 253 students, 118 teachers, and 3 administrators. A mixed approach was used: quantitative surveys (three adapted questionnaires) and in-depth interviews with management. AI penetration in education proved to be high: 32.7% to 58.3% of teachers and over half of students use generative AI. Substitution and augmentation (SAMR) dominate - AI is used to generate illustrations and tests, and less often for critical content analysis and project work. Key barriers are psychological (74.8%) and methodological (54.6%), not technological. Teacher competency deficits stem not from a lack of technical skills but from insufficient integrative knowledge (AI-TPACK) regarding the pedagogically sound use of AI in mentoring. A typology of educational organizations by AI readiness was developed: innovative leaders (high institutional support), motivated followers (strong student interest, moderate teacher activity), stable traditionalists (gradual adoption), and conservative laggards (low AI use). SAMR analysis shows that substitution and augmentation prevail, while transformative practices (modification and redefinition) remain rare. The typology enables differentiated AI implementation strategies. Key directions for AI integration into mentoring include routine task automation, analytical support, and personalized content creation. Integrating AI into pedagogical mentoring requires shifting focus from technical tools to developing AI-TPACK competencies and addressing educators' value-based attitudes. A differentiated approach based on institution type enhances implementation effectiveness. Promising directions include AI tutoring and intelligent learning systems that preserve the anthropocentric nature of mentoring. The authors declare no conflicts of interests.