Teacher Readiness for Deep Learning in Islamic Education: A Rasch Model Analysis of Challenges and Opportunities

The integration of deep learning in education has the potential to enhance pedagogical practices, personalized learning, and adaptive instruction. However, Islamic schools face unique challenges in adopting AI-driven educational models due to technological limitations, digital literacy disparities, and regulatory constraints. This study assesses the readiness of Islamic school teachers in Indonesia to implement deep learning-based curricula, analyzing knowledge, attitudes, barriers, and demographic influences on AI adoption. A structured questionnaire was administered to 1,120 teachers across madrasahs, pesantrens, and Islamic private schools, with data analyzed using the Rasch measurement model to ensure psychometric validity. Differential Item Functioning (DIF) analysis was conducted to examine variations in readiness across gender, age, education level, teaching experience, and ICT knowledge. The results reveal moderate teacher readiness, with significant gaps in deep learning comprehension and practical implementation. Female teachers, mid-career educators (36–45 years), and secondary school teachers exhibit higher AI readiness, while novice and older teachers face greater barriers. ICT literacy emerges as the strongest predictor of readiness, underscoring the need for targeted digital training programs. Findings highlight infrastructure deficits, professional development gaps, and policy misalignment as primary obstacles to deep learning adoption. While urban teachers demonstrate higher AI engagement, rural educators require greater institutional support. The study emphasizes the necessity of differentiated professional development programs that cater to teachers at different career stages and digital literacy levels. These insights provide critical implications for policymakers, educational leaders, and curriculum developers in designing AI-driven pedagogical strategies for Islamic schools. Future research should explore mentorship initiatives and hybrid training models to foster sustainable AI adoption in religious education settings.