An AI copied millions of images from manga and anime

The founder of Midjourney, a San Francisco-based artificial intelligence program, has admitted that he scraped some 100 million images from the internet, many of them without permission from official anime and manga creators.

Via Petapixel, Midjourney founder David Holz admitted in a resurfaced interview that the site has copied around 100 million images, revealing to Forbes how it was made. "It's just a big download from the internet. We use the open datasets that are published and train on them. And I would say it's something that 100% of people do. We weren't picky." He said they didn't seek consent from living artists or works that were still protected by copyright, adding, "No. There's really no way to get a hundred million images and know where they came from."

Holz's argument for why Midjourney did not seek consent stems from what he considers a lack of proper method of identification of which art belongs to whom. "It would be great if the images had embedded metadata about the copyright owner or something. But that doesn't exist; there is no record." As for the opt-out, he adds: "We are looking into it. The challenge now is to figure out what the rules are and how to find out if a person is really the artist of a particular work or just putting his name on it. We haven't found anyone who wants their name removed from the dataset that we've been able to find."

Although proving original ownership can be more difficult for smaller artists, the record of artists used by Midjourney includes high-profile and easily verifiable creators like Eiichiro Oda of "One Piece," Masashi Kishimoto of "Naruto," and thousands of other artists. Despite this, Midjourney users can generate AI art based on their works. Recently, anime association NAFCA sat down with Magmix and ethical AI developers Anime Chain to discuss what could be done in the face of the rise of AI in art creation. Anime Chain argued that AI was inevitable and that anime creators needed to take the lead before big tech monopolized the field.

Source: Forbes