Developing Pedagogical Innovation: AI-Based Deep Learning Training in Elementary Education

Abstract

Teacher readiness in addressing the challenges of 21st-century education requires a strong understanding of Deep Learning approaches as well as the ability to utilize artificial intelligence technology to enhance instructional quality. This community service program was designed to broaden teachers’ knowledge of the concepts and implementation of Deep Learning, which include meaningful, mindful and reflective, and joyful learning, while simultaneously equipping them with practical skills in the use of digital technology, particularly ChatGPT, as a learning support tool. The training was conducted on 27–28 October 2025 at SD IT Ar Rasyid Makassar with 25 participating teachers. The training methods included material delivery, group discussions, hands-on practice, case studies, and continued mentoring after the face-to-face sessions. Evaluation was carried out through pretests and post-tests, observations of participant engagement, and assessments of learning design outputs developed during the program. The evaluation results indicate improvements in teachers’ understanding of Deep Learning principles, the strengthening of 21st-century competencies, and enhanced ability to design context-based learning with artificial intelligence integration. An increase in the average score from 67.25 to 84.10 demonstrates that this program was effective in enhancing teachers’ capacity to implement more interactive, innovative, and learner-relevant instruction in the digital era. Therefore, this training provides a significant contribution to supporting learning transformation in schools.