Visual rhetoric and enunciation in the domain of multimodal generative artificial intelligence

Abstract

This article presents some proposals in visual rhetoric based on the contributions of the Group µ and Greimasian and post-Greimasian semiotics and tests them in the visual productions of generative AI. In the light of Greimasian and post-Greimasian semiotics, this article analyzes two emblematic cases that refer to media practices of image generation: an image that won a photography award and an image of an influencer that circulates on various social media platforms, such as Instagram. In our study, we use the concepts of plastic categories, figurativity, and enunciative praxis. We also draw on recent literature on generative artificial intelligence and, in particular, on word-to-image translation. Our goal is to understand how the Greimasian and post-Greimasian perspectives on rhetoric can help us analyze the phenomenon of generative artificial intelligence. The analysis of these two images revealed, on the one hand, some visible traces of the instauration process, and on the other hand, the instauration process can only be perceived through the record of the work of algorithms and it is not visible at the naked eye. Rhetoric provided us with resources to understand the phenomenon of image manipulation in its game of simulacra based on the relationship between truth, secrecy, and falsehood and the conflict in the enunciative contract between enunciator and enunciatee.