Epistemología de la imagen generativa: plausibilidad, estética y representación del pasado en imágenes producidas por inteligencia artificial

Abstract

This article examines the epistemological implications of artificial intelligence-generated images without empirical referents, focusing on their credibility, historical plausibility, and impact on collective memory. Drawing on visual culture and the philosophy of history, it argues that their plausibility stems not from documentary evidence but from the recognition of shared aesthetic codes. Methodologically, the study adopts a hermeneutic and critical approach, based on the analysis of selected cases of synthetic medieval representations disseminated on social media. Although entirely fictional, these images reproduce stylistic attributes established in historical painting and iconography, activating a collective aesthetic memory that renders them seemingly credible. The study concludes that their epistemic authority lies not in referential fidelity but in their capacity to mobilize culturally validated visual repertoires. This poses challenges for historiography, heritage transmission, and the uses of the past in the digital era, within a qualitative framework centered on image theory, philosophy of history, and software studies.