Abstract

This study examines the ethical issues arising from the integration of artificial intelligence (AI) technologies into Sīrah research and discusses the necessity of developing a comprehensive ethical framework in this context. With the advancement of digital humanities, classical sources on the life of the Prophet Muhammad have become analyzable through translation systems, natural language processing (NLP), machine learning (ML), simulation, and visualization techniques. However, this process poses serious epistemological and moral risks, such as the trivialization of the Prophet’s spiritual status, the neglect of his prophetic mission, and the reduction of the Sīrah to a mere historical narrative. In particular, it is emphasized that AI systems grounded in the epistemology of transhumanist philosophy may produce results that conflict with Sīrah, a discipline directly linked to divine revelation. Additionally, the study analyzes in detail the risks associated with digitizing classical Sīrah texts, including data representation issues, algorithmic bias, decontextualized translation errors, and sectarian tendencies. To address these problems, the study proposes technical solutions such as guardrail control algorithms, retrieval-augmented generation (RAG)-based transparency models, and Ara-TEaR-supported translation systems. Furthermore, it argues for the establishment of an AI ethics framework based on principles unique to the Sīrah discipline, in harmony with Islamic sources, prioritizing scientific and ethical oversight, and incorporating values such as preservation of the Prophet’s spiritual status, fairness in data representation, transparency in research models and methods, accountability, continuous human supervision, accessibility and inclusivity, hierarchical system instructions, and an academic citation regime for AI-generated content. The proposed framework aims to ensure the security of digital studies on the Prophet’s life and provide a systematic structure for the ethical use of AI in Sīrah research.