<![CDATA[Islamic Religious Education (PAI) plays a normative role in shaping students’ faith, piety, and moral character. However, its implementation remains largely cognitive, textual, and teacher-centered, limiting the development of students’ affective and spiritual dimensions. This study aims to reconstruct a deep learning-based PAI model to enhance students’ spiritual-reflective awareness in the Society 5.0 era. Using a mixed methods approach with a Design-Based Research (DBR) framework, the study was conducted through iterative stages: needs analysis, model design, implementation, evaluation, and revision. Data were collected through observations, interviews, questionnaires, and reflective journals. The findings indicate a significant transformation from surface to deep learning, evidenced by increased student engagement, reflective thinking, and spiritual awareness. Quantitative results show an improvement in overall scores from 3.05 (moderate) to 4.2 (high), while qualitative data reveal shifts in students’ perspectives from exam-oriented learning to meaningful spiritual understanding. The integration of experiential learning, reflective inquiry, and spiritual internalization proved effective in fostering self-reflection (muhasabah) and value internalization. Furthermore, the model demonstrates strong relevance to Society 5.0 by integrating technology with humanistic and spiritual values. This study concludes that deep learning-based PAI reconstruction is a strategic approach to developing holistic, reflective, and spiritually aware learners.]]>
