<![CDATA[Artificial intelligence (AI) has begun to reshape teachers’ professional work, including lesson planning, media design, assessment preparation, and collaborative reflection. However, the integration of AI in early childhood science learning remains pedagogically sensitive because young children require concrete, play-based, sensory, and teacher-mediated inquiry experiences rather than technology-driven instruction. This study examined how teacher-principal collaborative inquiry supported the contextual use of simple AI tools in early science learning at PAUD Sekar Nagari, Universitas Negeri Semarang. Using a qualitative case study design, data were collected from seven participants consisting of one principal, one PAUD coordinator, and five classroom teachers through open-ended online interviews, participant observation, and document analysis. The data were analyzed thematically through data condensation, data display, and conclusion drawing, with trustworthiness strengthened through triangulation, member checking, audit trails, and thick description. The findings reveal five interrelated themes: teachers’ cautious but positive initial perceptions of AI, transformational and digital leadership as an enabling condition, collaborative inquiry as AI-mediated professional learning, perceived improvement in teacher creativity and administrative efficiency, and continuing pedagogical-ethical boundaries in AI-assisted early science learning. Simple AI tools supported teachers in transforming science ideas into visual, concrete, and developmentally appropriate learning media, while collaborative inquiry helped them verify, adapt, and contextualize AI outputs. The principal’s leadership was decisive in creating psychological safety, facilitating infrastructure, and sustaining a professional learning community. The study concludes that AI can strengthen early science learning when positioned as a teacher-support tool rather than a substitute for hands-on inquiry, direct child observation, and teacher-child interaction.]]>
