Seeing is no longer believing: visual-AI, deepfakes, and the semiotics of synthetic images

In a world flooded with images, emerging visual technologies are playing a pivotal role in transforming how images are created, interpreted, and trusted in human vision. This article explores how machine-driven vision fits into a long history of mediated seeing, including drone photography. Drawing on a multimodal social semiotic framework, the study examines two widely circulated case studies: the deepfake of Pope Francis in a puffer jacket and the forged arrest of Donald Trump. Through this approach, the article demonstrates how such images derive from the mobilisation of familiar visual narratives. These synthetic visuals function as culturally resonant signs, blurring the boundaries between realism, satire, and misinformation. In response to these developments, the article suggests that visual culture is transitioning into a post-photographic epoch, characterised by declining reliance on images as direct evidence and their replacement by computer-generated representations. As AI-generated images circulate as seemingly real records, they challenge public trust in visual perception and highlight the need for enhanced visual, digital and algorithmic literacy. These literacies enable viewers to critically understand synthetic images and navigate complex visual environments where authenticity cannot be taken for granted.