Abstract

This article refers to the concept of “epistemia” as a diagnostic category for understanding the epistemological transformations brought about by the use of large language models and generative artificial intelligence. Far from constituting a theory of artificial intelligence or an ontology of cognition, epistemia describes a structural phenomenon emerging from sociotechnical systems: the production of linguistically coherent outputs that lack robust epistemic grounding. This paper argues that this phenomenon results from a deep compatibility between probabilistic text generation and human expectations of meaning, leading to a systematic confusion between plausibility and truth. Furthermore, epistemia is analyzed in relation to cognitive debt and cognitive offloading, highlighting how the externalization of intellectual tasks reshapes habits of judgment, attention, and understanding. A Platonic myth and a Borgesian tale are employed as hermeneutical devices that intensify the perception and analysis of the current phenomenon by offering formal structures that illuminate, by analogy, its ontological and epistemological implications. The analysis situates epistemia within broader debates in philosophy of science and technology, emphasizing its implications for the nature of knowledge, the limits of computational mimesis, and the future of human cognition in AI-mediated environments.