The persistent gap between the theoretical knowledge acquired at academic institutions and its practical application in the field remains a fundamental challenge in teacher training. The teacher educator serves as the critical mediator in this process and constitutes a vital bridge between theoretical knowledge and practical experience in the classroom (Fridman and Aaronson, 2025). By virtue of their role, teacher educator must possess a broad theoretical foundation and assist the student via a process of reflection and integrative connection between the experiences in the field and the pedagogical principles being taught (Shavit-Miller et al., 2022).

To facilitate an effective reflective process, teacher training often employs systematic frameworks, such as the Korthagen's ALACT model. This model requires the pre-service teacher to link classroom actions (stages 1–2) to an awareness of essential aspects, personal skills, and perceptions (stage 3), before formulating theory-based alternative actions (stage 4). However, for this process to be fully realized, there is a need for broad and immediate access to a variety of pedagogical and didactic theories that can illuminate the event from multiple perspectives (Korthagen, 2004).

The challenge lies in the fact that the breadth and depth of a teacher educator's theoretical knowledge is limited, as they are not necessarily an expert in every pedagogical discipline as many teacher educators have emerged from the education system after many years of teaching in a limited number of fields or disciplines. In this context, Orland-Barak's (2005) concept of 'lost in translation' is relevant to examining the lack of a dedicated academic background for the educational role; the lack of an appropriate theoretical foundation may limit the ability of these leaders to provide reflective feedback grounded in a wide range of perspectives and approaches. In recent years, there has been a growing understanding that the teacher educator, as someone entrusted with training the future generation of teachers, must himself undergo specialized professionalization and training processes. This training is seen as essential to fulfilling his role and aims to expand his teaching repertoire while deepening his professional identity. It is appropriate that these processes take place in a supportive environment close to the authentic teaching context, while relying on peer dialogue and analyzing examples from the field (Ullman-Darom & Rubin, 2022).

In recent years, there has been a growing trend in using Artificial Intelligence (AI) in teacher training.  AI can create classroom management simulations, assist in writing lesson plans, and improve the quality of feedback and reflection from pre-service teachers (Kumari, 2025). The use of chatbots based on generative Artificial Intelligence (GAI) in teacher training represents a profound change in the perception of the roles of the pre-service teacher and the teacher educator. Acting as a cognitive assistant, the chatbot enhances pre-service teachers' training through simulations and real-time theoretical application in lesson planning (Biberman-Shalev, 2025). It supplements the teacher educator's role by providing an extensive and adaptable knowledge resource. 

The aim of this research is to explore the impact of the integration of AI-based chatbot tools into the pedagogical feedback process of pre-service teachers as part of their training.

Research Questions

1. How does the integration of AI chatbot tools affect the theoretical grounding of pedagogical feedback compared to traditional feedback based on professional intuition?
2. In what ways does the use of chatbot tools expand the variety and scope of "action toolboxes" and strategies suggested to pre-service students during the feedback process?
3. To what extent does AI chatbot contribute to the structural consistency and objectivity of feedback?

This study employs a qualitative-interpretive approach to examine the influence of Generative AI on pedagogical feedback provided by teacher educators. The data set comprises feedback samples generated by 15 teacher educators at pre-service teacher training colleges. Data collection followed a dual-stage recursive process: • Phase 1: Human-Centric Feedback (Pre-AI): teacher educators documented their initial evaluations for students based on their existing disciplinary and pedagogical knowledge. These were captured via written reports or audio recordings. • Phase 2: AI-Enhanced Feedback: Leveraging Generative AI as a cognitive co-pilot, teacher educators refined and expanded their initial evaluations, resulting in a secondary, enhanced version of the feedback. Data was analyzed using Reflexive Thematic Analysis (RTA). The analytic process followed the six-phase framework originally proposed by Braun and Clarke (2006), while incorporating their updated conceptualization of the method as a reflexive, researcher-informed interpretive process (Braun & Clarke, 2021). Technological Development: The "Pedagogical Assistant" Chatbot. The study involved the development of a specialized chatbot based on the Gemini Large Language Model (LLM). Using Prompt Engineering techniques, the researchers designed Structured Prompts that defined the bot's persona as an expert pedagogical mentor. This configuration included a specific body of pedagogical knowledge and a reflective feedback style designed to offer scaffolding for creating in-depth, theory-based evaluations. Initial findings indicate a significant shift in instructional discourse. While original feedback (Phase 1) often focused on technical classroom management, the AI-enhanced feedback (Phase 2) demonstrated greater reflective depth and solid theoretical underpinning. Case Illustration: to illustrate the shift from technical description to theoretical analysis, consider the feedback provided to Student N: Human-only feedback: " You played music and did hand movements... I understood you wanted to make them concentrate". AI- Enhanced Feedback: Analyzes the same interaction through the lens of "Embodied Cognition theory": "When the pre-service teacher asked the students to make hand movements, she essentially "awakened" their nervous system. Physical activity helps with emotional regulation and prepares the brain for absorbing new information".

The analysis of the initial data set revealed three themes. First: Link between practice and theory: Feedback without the chatbot integration does not contain theories and explanations of theories. It mainly contains methods of action. Feedback tends to be more subjective and is based on "professional intuition" and experience. Feedback with the chatbot becomes more academic and grounded. Chat helps you pull out concepts and tie them directly to a specific moment in the lesson. This turns feedback into a learning tool for the student, not just an evaluator’s tool. In addition, the chatbot expands knowledge and allows for a deeper understanding of certain theories. Second: Level of detail and variety in action suggestions: Feedback without the chatbot: As humans, we tend to recommend strategies that we like or know well, suggesting a personal/subjective bias. Feedback may be repeated across students. Feedback with the chatbot: The chatbot can suggest 4-5 different courses of action for the same event. This expands the “toolbox” you offer the student. Third: Structure and Objectivity Feedback without the chatbot: Feedback can be influenced by the emotion of the moment or the personal relationship with the student or a bias. It may be less structurally organized. Feedback with the chatbot: The bot helps maintain a consistent structure (observation -> theoretical analysis -> suggestion for improvement). The feedback looks more professional and organized. Limitations of study The results presented here are preliminary and based on initial data, the full study presented at the conference will encompass a broader comparative analysis including feedback from a larger cohort of teacher educators.

Biberman-Shalev, L. (2025). Prompting Theory into Practice: Utilizing ChatGPT-4 in a Curriculum Planning Course. Education Sciences, 15(2), 196. https://doi.org/ 10.3390/educsci15020196 Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa Braun, V., & Clarke, V. (2021). Thematic analysis: A practical guide. SAGE Publications. Friedman, A., & Aaronson, M. (2025). Guidelines for the Profile of the Pedagogical Supervisor. Efrata Academic College of Education. Korthagen, F.A.J. (2004). In search of the essence of a good teacher: towards a more holistic approach in teacher education. Teaching and Teacher Education, 20(1), 77-97. https://doi.org/10.1016/j.tate.2003.10.002. Kumari, D. A., Begum, D. S., Paunikar, M. S., Kaur, A. & Verma, D. S. (2025). The Role of Artificial Intelligence in Teacher Training: Enhancing Pedagogical Effectiveness. Journal of Marketing & Social Research, 2(5), 116-122. Orland-Barak, L. (2005). Lost in translation: Mentors learning to participate in competing discourses of practice. Journal of Teacher Education, 56(4), 355–366. https://doi.org/10.1177/0022487105279077 Shavit-Miller, A., Rosenberg, K., Zuzovsky, R., Aldor, N., & Arviv-Elyashiv, R. (2022). Who are you, the pedagogical supervisor? Role perception, professional identity and shaping factors according to the perception of pedagogical supervisors at Kibbutzim College of Education [In Hebrew]. Kibbutzim College of Education, Ullman-Darom, R., & Rubin, Y. (2022). What else is there to innovate? Pedagogical training for new supervisors in physical education in elementary schools [In Hebrew]. Tel-Hai Academic College.