<![CDATA[This study investigates the effectiveness of an Artificial Intelligence-Supported Research-Based Learning (AI-RBL) model in strengthening the creative scientific competence of pre-service primary school teachers and examines how they perceived the role of AI during inquiry activities. A mixed-methods design with a one-group pre-test and post-test structure was implemented over a 15-week semester, during which a large language model acted as a digital mediator in problem formulation, hypothesis development, methodological planning, literature exploration, and data interpretation. Creative scientific competence was operationalized as originality, flexibility, elaboration, and relevance. These dimensions reflect key inquiry practices that underpin scientific literacy. Specifically, they involve the ability to generate meaningful questions, apply scientific reasoning, use evidence appropriately, and construct coherent explanations. A repeated-measures multivariate analysis of variance showed a significant effect of time, with improvements across all dimensions, and univariate tests indicated the strongest gains in elaboration and flexibility. Qualitative reflections revealed that the model supported divergent idea generation, strengthened structural reasoning, and encouraged the critical evaluation of AI suggestions, thereby enhancing both cognitive and metacognitive aspects of inquiry-based learning. The findings highlight the educational significance of pedagogically constrained AI mediation in developing inquiry-oriented competencies that underpin scientific literacy, and they contribute to international discussions on responsible and pedagogically grounded integration of AI in science teacher education.]]>
