Bradley C. Dart

This paper investigates the epistemic and moral untrustworthiness of artificial intelligence (AI) as a mediator of Islamic knowledge in an online environment characterized by misinformation, polarization, and systemic bias. The structure of digital media – shaped by cognitive biases, network centralization, and profit-driven algorithmic design – amplifies distortions of truth through mechanisms of engagement, surveillance, and media logic. These processes generate an attention economy that rewards divisive or sensational content, fostering echo chambers and reinforcing societal prejudices. When this system is coupled with machine learning models trained on biased data, it produces AI outputs that replicate Islamophobic associations and propagate misinformation about Muslims. The paper distinguishes between epistemic trust, appropriate for limited cognitive functions of AI, and moral or religious trust, which requires virtue, intention, and accountability – qualities AI inherently lacks. Using the Islamic tradition’s emphasis on personal transmission of knowledge, moral character, and verified scholarly authority, the study contrasts interpersonal chains of trust with algorithmic systems governed by opaque data and economic incentives. Consequently, while AI may serve as an epistemic tool for accessing vetted information, it is unfit to function as a moral or religious authority. The paper concludes by advocating for enhanced digital literacy, algorithmic transparency, community-based factchecking, and the development of an Islamically grounded ethics of technology. Such measures aim to mitigate epistemic harm, resist algorithmic authority in religious matters, and reaffirm the centrality of human scholarship and divine revelation in the transmission of Islamic knowledge.