Open Collections

Abstract

Contemporary artificial intelligence systems excel in controlled environments but struggle to navigate uncertainty and contextual complexity that characterize real-world decision-making. This limitation stems from the rigidity of AI models and insufficient contextualization, reducing their ability to adapt beyond narrow optimization tasks. While AI researchers experiment with technical architectures for improved decision-making, there has been little investigation into how classical legal systems might inform AI models capable of sound and adaptive reasoning. This thesis addresses this interdisciplinary gap by examining how Ḥanafī legal methodology can inspire more robust and ethically grounded artificial intelligence systems. Using hermeneutic textual analysis informed by Gadamer's fusion of horizons and Latour's Actor-Network Theory, this study analyzes classical Ḥanafī legal sources and contemporary scholarship alongside foundational AI literature. The research focuses on two key mechanisms: the qaṭʿī/ẓannī epistemological framework for evaluating certainty and managing uncertainty, and istiḥsān as a method for principled departure from precedent when rigid application yields inequitable outcomes. Through case studies of medical AI diagnostics, the thesis demonstrates how these classical Islamic legal methodologies address current limitations in algorithmic rigidity and uncertainty quantification. The analysis reveals that effective AI decision-making requires systematic integration of authoritative grounding with contextual adaptability. Three operational principles emerge: Authority Through Validation, requiring comprehensive evaluation of source reliability and interpretive clarity; Adaptability Through Guardrails, providing structured mechanisms for principled departure from standard processing; and Epistemic Humility, enabling transparent communication of confidence levels. iii This research contributes to AI ethics discourse by proposing epistemological transfer learning that draws insights from centuries of Islamic jurisprudential reasoning. The findings suggest that classical Ḥanafī legal methodology offers systematic frameworks for addressing authority- adaptability tensions that contemporary AI systems struggle to resolve, demonstrating the continued relevance of traditional Islamic interpretive frameworks for modern technological applications in healthcare artificial intelligence, machine learning ethics, and beyond.