Text-to-image technologies. The Aesthetic implications of AI-generated images.
Year: 2024
Type: article
Abstract: Synthography, a term that describes images generated by Text-to-Image (TTI) technologies such as DALL·E, Midjourney, and Stable Diffusion, enables the creation of images generated through software that involve linguistic prompts. Such prompts are processed by encoded semantic systems, able to capture compositional aspects of arbitrary language text inputs. By drawing on the emerging aesthetic theories on AI-generated images, this contribution aims at discussing the European AI-Act, published in March 2024. Firstly, it analyzes the Act's perspective on synthographies, highlighting how it interprets generative AI within the paradigm of copying and technical reproducibility. Secondly, the paper discusses the regulatory challenges posed by AI-generated content, which blurs the line between original works and replicas, raising concerns about copyright and the authenticity of creative outputs. Third, the study explores the Act's paragraph regarding AI systems that produce content resembling existing images, as it concerns the testimonial value of images. Finally, this contribution examines the possible contribution offered by AI-generated images in the field of the visual art.
Source: Itinera
Author: Lorenzo Manera
Institution: University of Modena and Reggio Emilia
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4406153366