- Vogue Italia magazine first published images created by artificial intelligence (AI) in its May 2023 issue
- These images show model Bella Hadid wearing fashion clothes and accessories and on backgrounds designed by Dall-E2 (OpenAI) and Midjourney
- This project aims to explore the creative possibilities of AI for a magazine like Vogue, but it also questions the notions of authenticity and representation in the field of fashion.
Vogue Italia magazine is causing a stir by publishing images created by an artificial intelligence (AI) for the first time in its May 2023 edition. These images show virtual models wearing clothes and fashion accessories designed by AI from visual data.
The project, called “Bella and AI”, is the result of a collaboration between the magazine and photographer Carlijn Jacob (assisted by AI specialist Chad Nelson). For the generation of images they used the AI DALL-E2, and Midjourney, both able to generate images from texts or keywords, but in a different way.
When AI invites itself into Vogue
Bella hadid for vogue italia may 2023 issue photographed by carlijin jacobs pic.twitter.com/p5GDvoBBmE
— it girl (@modelsdiaryy) April 27, 2023
Indeed, with Midjourney, all you have to do is enter a specific prompt and four images are generated. Images that can then be upscaled or remix (re-compose) to some extent (on the same basic idea). On the other hand, Dall-E works more like conversations with ChatGPT. We can say for example “add a palm tree in the background behind the subject” after the image has been generated.
The photographer continues: “After having crafted a rough concept for each photo, we photographed Bella in the studio and then tried to generate background images with the AI program”. And Carlsijn Jacobs later added: “As we used the program, I realized how difficult it was to entrust your vision to an interpreter. Despite Chad’s extensive experience with the instrument, DALL-E’s performance depended on the level of accuracy of the prompts, and finding the right ones to bring my particular vision to life proved difficult and frustrating on more than one occasion” .
The result that we invite you to discover in the source of this article and via the Tweet above, is in any case very surprising and confusing. The images of the shoot create a particular atmosphere, with sometimes the AI interpreting dresses from great designers like JW Anderson to better extend its characteristics into the set – without the need to create an expensive hard composition for the session. photo, a long and restrictive process to put in place.
“This experience has taught us that using AI to generate high-end imagery is always a daunting undertaking. Technology is, however, a valuable tool for quickly building mood boards and testing the most outlandish ideas more quickly”concludes Imruh Asha, a Parisian stylist who also participated in the project alongside Carlijn Jacobs.
The “Bella and AI” project raises questions about notions of authenticity and representation in fashion. How to distinguish the true from the false? What is the role of the human creator in relation to the AI? What is the artistic and commercial value of AI-generated images? How to protect copyright and intellectual property? So many questions that ask us, in the end, about what remains (and will remain in the future) that is truly unique and exclusive to human (artists) in the face of AIs that are increasingly showing surprising talents…
