<![CDATA[The emergence of Generative Artificial Intelligence (AI) has transformed contemporary knowledge practices by enabling machines to produce texts, explanations, and interpretations that increasingly resemble human intellectual activities. However, the ability of Large Language Models (LLMs) to generate coherent and convincing responses raises a fundamental epistemological question: does the production of knowledge-like representations indicate genuine knowledge possession, or does it merely constitute a simulation of knowing? This study aims to critically examine the epistemological status of Generative AI by investigating the distinction between knowledge, information generation, and simulated knowledge through the perspective of Al-Ghazali’s epistemology. Employing a qualitative conceptual approach, this research combines philosophical analysis of Generative AI, studies on AI hallucination and epistemic authority, and Al-Ghazali’s concepts of knowledge, doubt, and certainty as an analytical framework. The study reveals three main findings: first, Generative AI functions as a representational system that reconstructs patterns of information rather than as an epistemic subject capable of understanding and knowing; second, AI hallucination represents an epistemic illusion that exploits human vulnerability by masking linguistic fluency with the non-transferable state of genuine certainty (yaqīn); and third, moving beyond generic ethical clichés, the Ghazalian framework furnishes a concrete normative typology that categorizes uncritical reliance as blameworthy imitation (taqlīd) and elevates critical examination (naẓar) into a binding duty distributed between individual (farḍ 'ayn) and collective (farḍ kifāyah) obligations. This study concludes that integrating Islamic epistemology into contemporary AI discourse offers a rigorous philosophical architecture that positions AI strictly as an instrumental simulation engine while securing human intellectual agency.]]>
