Image generators take on the challenge of ugliness

If the aesthetic aspect of images created by artificial intelligence has now reached the ultimate stage of photorealism, their semantic field remains human, all too human.

Like the photographic process before it, prompt art - which consists of generating images by formulating text commands on an artificial intelligence (AI) interface - can now boast results that come close to the illusion of reality. While photography aims to offer a representation of reality through the mediation of an operator and his camera, its referent is inscribed in the materiality of the physical world; the computational method, on the other hand, generates a composite image devoid of this direct anchoring link, on the heterogeneous basis of billions of pre-existing visual references compiled on the Internet. In other words