Former US President Donald Trump struggling with the police, Pope Francis sporting a white down jacket, Emmanuel Macron in the middle of a demonstration… While images created by artificial intelligence (AI) have been circulating on the Web for months, these last snaps, born from the most recent version of the Midjourney image-generating software, have troubled.
Like the stunning texts of the ChatGPT software, the realism of these admittedly fictitious photos struck the general public. It also created a shock wave in the image world, now worried about the future of the work of photographers and illustrators.
That press sites took over this tool did not reassure them. Le Figaro for example, on March 26, used Midjourney for the illustration of a Web article, before going back the next day and replacing it with a photo.
Opposite Strategies
The monthly So Foot he assumes to use the default software, for lack of financial means, if he does not find a satisfactory photo. “Out of 6,000 content published since November [2022]we used Midjourney a dozen times, specifying it”relativizes Pierre Maturana, editorial director.
“We are a minimedia, with minimeans”abounds Pablo Pillaud-Vivien, the editor-in-chief of the magazine Regards, which also uses AI. The daily suisse Blick also took the plunge, on March 25, stating that“None of the people in this AI-generated photo exist”.
The trend also affects communication and advertising: the “one” of the latest Recrea Magthe Grenoble cultural diary, published in October 2022, was “designed” thanks to Midjourney, as was the poster for Nutcracker of the San Francisco Ballet, at the end of 2022. And, at the beginning of March, Coca-Cola invited Internet users to use Dall-E 2 – software developed by OpenAI, the creator of ChatGPT, and competitor of Midjourney – to generate creations, the best of which will be used by the brand.
Others have, conversely, already set clear limits: the daily Release will not use “images generated by AI, except for subjects directly related to AI”. A line close to that of Monde, where the publication of news images produced by AIs is also prohibited. THE New York Times has indicated in a red banner that the illustration of an article on AI was… “generated by an AI”.
In the powerful image banks, opposing strategies are also taking shape. Getty Images announced in September 2022 that it refused to store photographs created with AI, while Shutterstock entered into a partnership with OpenAI in 2021: the image bank allows Dall-E software to be trained on its database and provides its customers with the image generator.
You have 55.56% of this article left to read. The following is for subscribers only.