A Jurisprudential Reexamination of the Concept of Intention (Niyyah) and the Attribution of Responsibility in AI-Based Decisions: A Motivation-Oriented Approach

Abstract

The advent of artificial intelligence into sensitive decision-making arenas has transformed the very nature of decision and intention from a purely human act into a complex, technology-driven process. This transformation poses fundamental challenges to the Islamic jurisprudential (fiqh) system, particularly with regard to the "attribution of responsibility" (isnād-e mas'ūliyyat). This is because Islamic jurisprudence grounds responsibility on "intention" (niyyah) and "purpose" (qasd); however, the opaque structure (the "black box") of deep learning algorithms makes tracing the agent's intention exceedingly difficult. Unlike previous studies, which have largely focused on compensation and civil liability (ḍamān), this research employs a descriptive-analytical method and adopts a "neo-fiqh" approach to reexamine the concepts of purpose, motive, and direct agency (mubāshirat) within the technical context of AI.

The findings reveal that in the digital ecosystem, the traditional concept of "the agent's intention" has transformed into a "distributed intention" shared among the designer, the data trainer, and the user. Furthermore, in confronting algorithmic uncertainty, the psychological element of responsibility shifts in nature from "intention toward a specific outcome" to "intention to act and acceptance of risk" (grounded in the juridical rule of "action" or qā'idah-ye eqdām). This paper endeavors to move AI beyond its conventional status as a mere "tool" by distinguishing between the system's "technical agency" and the human's "jurisprudential agency" (fā'eliyyat), thereby proposing a novel model for attributing responsibility based on the "degree of participation in motive and intention." This model has the potential to fill the existing theoretical gaps in legislation concerning autonomous systems and to pave the way for the formulation of more precise legal frameworks.