Designing an AI-integrated role-rotation pedagogical model to support competence-related learning in pre-service educational psychologists

Synthesis: Kenzhebayeva, Matayev, Sarsembayeva, Kolyukh, Assenova, and Ospanova (2026) report a design-based research project that treated systematic role rotation as the organizing structure for critical human-AI interaction. Sixty-two undergraduates preparing as educational psychologists in Kazakhstan worked through gifted-education cases over eight weeks, moving successively through four professional positions: Case Constructor, Research Analyst, Practitioner-Interventionist, and Reflective Researcher, while using generative AI as a source of preliminary ideas. Across reflective journals, interviews, observational notes, and student-generated learning products, the authors describe analytical reasoning, contextual adaptation, intervention planning, and increasingly critical handling of generated recommendations, with AI framed as cognitive support rather than a substitute for professional judgment. Uneven participation, difficulty adopting unfamiliar roles, and occasional overreliance on AI recommendations also appeared across the dataset. Because the study collected no pre-post competence measures and used no comparison group, the authors present the results as context-specific evidence of engagement rather than demonstrated gains in professional competence.

Key Findings

• Role rotation, not AI use by itself, was the central pedagogical mechanism: four roles distributed problem framing, evidence evaluation, intervention planning, and reflective evaluation across successive case cycles.
• Sixty-two undergraduate pre-service educational psychologists in a Pedagogy and Psychology program at a Kazakhstani university completed an eight-week professional preparation intervention built on gifted-education cases and generative AI support.
• Data came from four complementary sources: reflective journals, semi-structured interviews, observational field notes, and student-generated learning products, analyzed with reflexive thematic analysis, triangulation, and member checking.
• Five themes captured participants' experience, spanning cognitive, practical, reflective, technological, and role-based dimensions of competence-related learning as reported across journals, interviews, observations, and learning products.
• Participants treated AI as a source of preliminary ideas and cognitive support, comparing its outputs with psychological theories and modifying or rejecting recommendations that did not fit case context.
• Persistent problems included overreliance on apparently authoritative AI responses, uneven participation, difficulty adapting to unfamiliar roles, and initial uncertainty during reflective and evaluative tasks in the first cycles.
• The authors caution that without pre-post competence measures or a comparison group, the findings show engagement during the intervention rather than measurable developmental gains, and they call for comparisons between rotating and stable roles.

The AI-Integrated Role-Rotation Model

The model combines four elements that the authors argue had previously been studied separately: AI-supported case work, case-based professional learning, collaborative role rotation, and structured reflection. Its conceptual base draws on Vygotsky's socio-constructivist account of learning through social interaction, Schön's reflective practitioner, and Kolb's experiential learning cycle. Participants did not hold a stable role. They moved successively through Case Constructor (contextual problem framing), Research Analyst (evidence evaluation and comparison with theory), Practitioner-Interventionist (contextual adaptation and justification of strategies), and Reflective Researcher (metacognition and critical examination of both AI outputs and peer decisions). The design conjecture was that this movement would create repeated opportunities for professional reasoning and critique of generated recommendations, since each role frames the same case differently. Role rotation is presented as a mechanism for structuring critical human-AI interaction in higher education, not as a participation strategy alone.

Study Design and Data

The study used design-based research across five iterative phases: problem analysis, pedagogical design, implementation, evaluation, and refinement. Participants were 62 second- to fourth-year undergraduates enrolled in a Pedagogy and Psychology program at Margulan University in Pavlodar, Kazakhstan, selected purposively because they already had foundational training in psychological and pedagogical support. Data collection combined reflective journals, semi-structured interviews, observational field notes, and student learning products, examined through reflexive thematic analysis and triangulation, with member checking, peer debriefing, and an audit trail supporting trustworthiness. Cognitive, practical, and reflective dimensions were operationalized as sensitizing concepts rather than as mutually exclusive categories. Ethical safeguards were explicit: all cases were fictional or fully anonymized, participants gave informed consent, participation was voluntary, and no identifiable educational or psychological information could be entered into AI platforms.

What Role Rotation Changed in Practice

Participants described each role as demanding a distinct way of thinking. The Case Constructor considered contextual factors before analysis; the Research Analyst compared AI suggestions with theoretical knowledge; the Practitioner-Interventionist adapted plans to specific circumstances; the Reflective Researcher examined the limits of both generated outputs and group decisions. Later-cycle observations recorded more requests for theoretical justification and more debate about contextual fit, and several learning products combined analytical evidence, practical recommendations, ethics, and evaluation in one plan. The two most consistent frictions were the Reflective Researcher role, which students found hardest because it required judging peers as well as AI, and unequal airtime, since more confident participants sometimes dominated discussion despite the assigned role structure. Some students also continued to seek confirmation from AI before offering their own interpretation, even in later cycles.

What this means for practice

• Teacher educators can treat the four roles as the load-bearing structure and AI as a support inside it rather than the center of the course.
• Start from authentic professional cases, since realistic scenarios gave participants something concrete to reason about and to check against theory.
• Then assign the rotation deliberately, so every student moves through problem framing, evidence evaluation, intervention planning, and reflection within a single course cycle instead of settling into one comfortable position.
• Make AI outputs the object of discussion: ask groups to verify generated recommendations against sources, request theoretical justification before adopting a plan, and note where a plan was modified or rejected.
• Add explicit prompts and facilitation for the Reflective Researcher role, which students found hardest, and monitor airtime so confident voices do not override the assigned structure. Revise role descriptions and reflection prompts between iterations. These moves assume Critical Thinking and Metacognition are taught rather than presumed.

Limitations

• The authors are explicit that the intervention ran in one university over eight weeks, with neither pre-post competence measures nor a comparison group.
• The patterns therefore document engagement in competence-related learning, not developmental gains.
• Four design principles are offered for future work: explicit role guidance and facilitation, deliberate exposure to multiple professional perspectives, structured collaborative reflection on AI-supported information, and authentic professional cases as the context for AI-supported activity.
• The model was adapted iteratively during the intervention, with role descriptions and reflection prompts revised as difficulties surfaced. The authors call for multi-institutional, longitudinal, and mixed-methods studies, and for comparisons between rotating and stable role structures, to test whether rotation adds distinctive benefits for perspective-taking, collaborative accountability, and critical evaluation of AI outputs.

Connected Concepts

• Generative AI — AI used as a source of preliminary ideas and cognitive support in case work
• Collaborative Learning — role rotation as a way to distribute responsibility across a group
• Metacognition — the reflective dimension of the Reflective Researcher role
• Critical Thinking — verification and contextual judgment applied to AI-generated recommendations
• Trust Calibration — overreliance on apparently authoritative AI responses
• Human AI Collaboration — critical human-AI interaction as the object of the design
• Design-Based Research — the five-phase iterative methodology

Connected Articles

• The Architecture of Roles in AI-Designed Group Activities: A comparative inductive analysis of novice and experienced teachers' differentiated instruction within the IAT framework — The Architecture of Roles in AI-Designed Group Activities
• The Reflective Triangle Model: AI as a Cognitive Mediator in Teachers' Professional Learning and Learning-Community Development — The Reflective Triangle Model: AI as a Cognitive Mediator
• Development and evaluation of artificial intelligence literacy training for teacher education students — GenAI Literacy Training for Teacher Education Students

Citation

Kenzhebayeva, Z., Matayev, B., Sarsembayeva, E., Kolyukh, O., Assenova, N., & Ospanova, A. (2026). Designing an AI-integrated role-rotation pedagogical model to support competence-related learning in pre-service educational psychologists.