Photo 265588290 © Rokas Tenys | Dreamstime.com
Midjourney once again faces scrutiny over its training methods no thanks to a circulating list that includes the names of 16,000 artists, whose work is believed to have been scraped to develop the popular text-to-image engine.
Over New Year’s weekend, a Google Sheet began circulating on social platforms X (formerly Twitter) and Bluesky, purporting to show the extensive database of artists whose work allegedly informed Midjourney’s artificial intelligence. This roster, known as Exhibit J, spans 24 pages and includes a wide array of artists, from iconic figures like Banksy, Anish Kapoor, Gustav Klimt, Hayao Miyazaki, Damien Hirst, Keith Haring, Matt Groening, and Yayoi Kusama to commercial illustrators and entities such as Walt Disney, Tim Burton, MSCHF, and Sony PlayStation.
A startling inclusion flagged up by ARTnews is that of six-year-old artist Hyan Tran, known for contributing to a fundraiser for Seattle Children’s Hospital in 2021
The original Google file has since been restricted, but a copy remains accessible on the Internet Archive.
The document is part of a larger class-action lawsuit targeting Stability AI, Midjourney, and DeviantArt last year, and has emerged as a November 29 amendment to the complaint after a California federal court judge dismissed several claims from a group of artists against Midjourney and DeviantArt. The case continues to unfold, reflecting the growing concerns and debates over the use and rights of creative content in training AI tools.
At the heart of this issue is the broader conversation about copyright, consent, and compensation in the digital age, especially as it pertains to AI's rapidly evolving capabilities. As the industry pushes forward, the dialogue surrounding the ethical use of artists’ work becomes increasingly urgent.
[via ARTnews, images via various sources]