Constructing Digital Religious Agency: Generative AI Literacy and Islamic Learning within the Merdeka Curriculum Framework

The increasing integration of generative artificial intelligence into educational practices has introduced new challenges for student agency, particularly in subjects involving normative and epistemic authority such as Islamic Religious Education. While the Merdeka Curriculum emphasizes learner autonomy, limited attention has been given to how this autonomy operates when students increasingly depend on AI-generated religious explanations. This study investigates how Madrasah Aliyah students construct digital religious agency through interactions with generative artificial intelligence in Islamic learning contexts. This study adopts a qualitative design using Interpretative Phenomenological Analysis. Data were collected through in-depth interviews, focus group discussions, and digital artefact analysis involving fifteen students and three Islamic Religious Education teachers in Bogor, Indonesia. Findings identify three typologies of digital religious agency. First, pragmatic-instrumental agency emphasizes efficiency-oriented use of AI. Second, explorative-critical agency involves verification, comparison, and reflective engagement. Third, passive-acceptive agency reflects uncritical reliance on AI-generated knowledge. Results show that although the Merdeka Curriculum legitimizes autonomous engagement with digital resources, the absence of structured guidance on artificial intelligence literacy often leads to superficial autonomy and epistemic dependence rather than reflective independence. Student agency in AI-mediated religious learning should therefore be understood as a multidimensional capacity encompassing technical competence, epistemic awareness, and evaluative judgment. Integrating critical artificial intelligence literacy into Islamic Religious Education is essential to ensure learner autonomy contributes to responsible and informed religious understanding in the digital era.