This study explored the lived experiences of junior high school science teachers in overcoming challenges related to the integration of artificial intelligence (AI) in classroom instruction. Anchored in a qualitative phenomenological design, the research involved eleven (11) purposively selected teachers from a public secondary school in Leyte Division. Data were collected through open-ended questionnaire responses and analyzed using the phenomenological approach of Clark Moustakas (1994), focusing on identifying significant statements, clustering themes, and deriving the essence of participants’ experiences. Findings revealed that teachers exhibit varying levels of AI competence, ranging from developing to skillful utilization, with many prioritizing practical applications over technical expertise. Key challenges identified include limited technological resources, concerns about the quality and reliability of AI outputs, and issues related to academic integrity. Additionally, insufficient training and lack of hands-on experience hinder effective AI integration. Participants emphasized the need for continuous, practical training, collaborative learning environments, and access to advanced technological resources. The study highlighted the gap between policy expectations and actual classroom practices, underscoring the importance of sustained professional development and institutional support to enhance teachers’ competence and confidence in integrating AI into science education.