Islamic Religious Education (IRE) plays a central role in shaping the character and morals of students at various levels of education. However, challenges arise when IRE policies are not fully able to adapt to technological developments and the demands of 21st-century learning. This study aims to analyze the application of the deep learning approach as an innovation in improving the quality of PAI policies. The research method used is a qualitative study based on literature analysis to explore the potential, opportunities, and obstacles to implementing deep learning in the context of religious education. The results show that the integration of deep learning can improve the personalization of the learning process, strengthen data-based evaluation, and encourage more adaptive and responsive policy effectiveness. However, the application of this technology faces a number of obstacles, including limited digital infrastructure, low teacher technological competence, and ethical issues surrounding data use. The implications of this study emphasize the need for flexible PAI policies and comprehensive teacher training programs so that the application of deep learning can run optimally. Thus, this study provides a conceptual contribution to the transformation of religious education policies towards a system that is of higher quality, sustainable, and relevant to the digital era.
Copyrights © 2025