David Holz, the founder of Midjourney, recently admitted something we’ve already assumed: the company’s AI was trained on hundreds of millions of images without consent from their authors. This revelation has sparked outrage among both artists and privacy advocates. It has raised concerns about the ethical implications of such actions, as well as copyright issues that might emerge.
Back in September, Holz gave in an interview to Forbes, revealing that his company didn’t seek consent from artists when using their work for AI training. He justified it with the lack of possibility to “get a hundred million images and know where they’re coming from.”
“It would be cool if images had metadata embedded in them about the copyright owner or something. But that’s not a thing; there’s not a registry. There’s no way to find a picture on the Internet, and then automatically trace it to an owner and then have any way of doing anything to authenticate it.”
words of David Holz (midjourney founder), from forbes article (link below): pic.twitter.com/rnWP28rrag
— Maciej Kuciara (@maciej_kuciara) December 20, 2022
If this makes you furious, wait, there’s more. Forbes asked Holz if artists can opt out of being included in Midjourney’s data training model, or from being named in prompts. Can you guess his answers? You’re right, it’s a hard no on both, at least for now. Holz said that the company is “looking at that,” adding that “the challenge now is finding out what the rules are, and how to figure out if a person is really the artist of a particular work or just putting their name on it.” The only thing you can do for now is to check whether your art was used to train AI, and this feature has nothing to do with Midjourney, it is a tool on its own.
Forbes discussed the future of art that’s certainly already being impacted by Midjourney and other AI image generators. Will AI art destroy artists’ livelihoods? Holz sees two possible scenarios:
“I think there’s kind of two ways this could go. One way is to try to provide the same level of content that people consume at a lower price, right? And the other way to go about it is to build wildly better content at the prices that we’re already willing to spend. I find that most people, if they’re already spending money, and you have the choice between wildly better content or cheaper content, actually choose wildly better content. The market has already established a price that people are willing to pay.”
I agree up to a point: I’d always rather pay a living being, an artist, to create something of great quality. However, many people would rather pay less money for mediocre or under-average work than pay a higher price for high-quality work. There are still those willing to pay for excellent work – but I feel like their number is decreasing. And what are we going to do if AI-generated art becomes better, as I’m sure it will? In this case, there’s double damage.
First, all those artists whose work was used for AI training will stay without compensation for having their work contribute to the development of these tools. And second, many more artists won’t get gigs even if they undercharge for their work, as there will always be an AI tool ready to do it for free, or for even less cash.
While I’m excited and intrigued to see text-to-image and other AI technology expand, I’m trying to curb my enthusiasm and stay aware of the potential harm that it could do. I’m not saying there aren’t good sides to Midjourney and other text-to-image generators and I sure enjoy playing around with them. But using a hundred million images without consent isn’t one of the good sides, I’d say. it opens up a Pandora’s Box of various copyright issues, and we’re still to discover what will get out of it.
[via PetaPixel]
FIND THIS INTERESTING? SHARE IT WITH YOUR FRIENDS!