<![CDATA[Background: The rapid growth of educational technology demands teachers’ abilities to integrate AI into learning practices. Yet, many educators still struggle to design interactive and constructive media aligned with pedagogical goals.Aims: The goal of this community service program was to empower lecturers, teachers, and students in Bengkulu University's Physics Education Study Program by developing interactive and constructive learning media that were integrated with AI. The activity was carried out using a participatory and collaborative training paradigm with four stages: preparation, implementation, mentorship, and evaluation.Methods: The methodology framework used a descriptive qualitative approach with pre-test and post-tests, observations, and participant reflections. Training materials were developed using the Technological Pedagogical Content Knowledge (TPACK) framework, which emphasizes the integration of pedagogical and technological skills. Participants took part in hands-on workshops where they designed physics learning media prototypes utilizing AI-based platforms like ChatGPT, Canva, and Genially.Results: The findings revealed a considerable boost in participants' competencies, with post-test scores improving by 35% on average. Qualitative data indicated high levels of satisfaction (96.4%) and considerable appreciation for the practical sessions and coaching. Participants showed improved proficiency in aligning AI-assisted technologies with constructivist learning concepts and ethical awareness. The program also built a digital community of practice to ensure sustainability and collaboration after the training term. Conclusions: The findings show that incorporating AI-based learning media into community service can effectively promote teacher creativity, digital competence, and responsible technology usage in a period of educational transition.]]>
