Large Language Models (LLMs) as Islamic Guidance Tools: Trust, Limitations, and Ethical Boundaries

الملخص

The surge in the availability of large language models (LLMs) has opened up vast prospects and at the same time posed major challenges for Islamic guidance.This study evaluates the role of LLMs in the provision of religious advice for Islam in a critical way. The trust relationships between users, AI, and traditional authorities who are the worm of offense and the peaceable kingdom of their existence and how these relationships are established in the context of LLMs are discussed here. The study makes use of a qualitative, text-based, and critical analysis of original Islamic legal sources, general references, and written LLMs interactions to draw a comparison between LLMs capacities and conventional Islamic epistemological understandings. Results indicate the possibilities of LLMs giving access to the democratization of Islamic information at various levels, allowing for the automation of answers to frequently asked questions, and contributing to the educational networks. At the same time, there are certain critical areas that prove to be stumbling blocks, disabling the otherwise promising idea. In these areas, theological worries about the misunderstanding of sacred texts, data-bias and absence of context adaptation in the technology due to advances, which include technical limitations, and firasah, that the researchers mentioned, the negative effect of the power needed to uphold scholarly authority. There is a huge and important issue which is the privacy factor and the commercialization issue can be brought up to the light as the vast and detailed information might be accessed and/or used commercially. The key to whether the guidance that is derived from LLM is trustworthy comes from the fact that it is transparent and the accuracy of the answers can be attested to by the authoritative sources (Quran, Sunnah) alongside it being under the continuous surveillance of the qualified Islamic scholars.