Disputed Memories: Generative AI Images and Visual Reconstructions of the Past

Abstract

This article analyzes how generative AI images participate in a visual reconstruction of collective memory and in symbolic disputes surrounding the past. This research is justified by the increasing circulation of synthetic images in digital environments and by the need to understand their effects on contemporary ways of remembering, imagining, and representing historical experiences. The study draws on a dialogue between memory studies, image theory, AI critique, and techno diversity, understanding technology as a cultural, political, and situated phenomenon. Methodologically, a qualitative, exploratory approach is developed in two stages: AI-generated, wholly or partially, visual production mapping, and an interpretive analysis of a corpus comprising the works of Brunno Sarttori, Mayara Ferrão, Pedro Garcia, and André Lima, as well as Orixás da IA’s profile. The analysis articulates three categories: remembering/celebrating, past imaginaries, and resignification, understood as mediative gestures between image, memory, and technology. Results indicate that these productions not only recover nostalgic or historical references but also elaborate possible, affective, speculative pasts, especially in narratives linked to historically marginalized groups. It concludes that generative AI constitutes an ambivalent field of dispute over memory. At the same time, it can reproduce hegemonic patterns and cultural biases; it can also be appropriated to imagine other presences, visualities, and forms of historical inscription.