A class action lawsuit has been filed in California against Stability AI, Midjourney and DeviantArt by artists who allege that these companies have infringed their copyright. The problem stems largely from the use of copyrighted images in massive data catalogs used to train these companies’ AI art generation tools, which are often used without artist consent or compensation of any kind.
The artists are represented by the law firm Joseph Saveri (via Ars Technica (opens in new tab)), who notes, “Stable Diffusion and other image-generating AI products could not exist without the work of painters, illustrators, photographers, sculptors and other artists.”
“…The creators of these image generators have violated the rights of at least thousands of artists when they created their various products. Although proponents of AI-generated images have argued that the creation and use of these products is perfectly legal, no court has yet answered the question.”
It’s true, AI art generation tools like Stable Diffusion and Midjourney rely on huge datasets to train their algorithms to produce new images based on human-entered prompts. Without training, they wouldn’t have any knowledge of what an art style is, what objects are, or any information needed to do their job.
The thing is, those datasets are often made from copyrighted material. A company LAION (opens in new tab) sidesteps copyright claims by noting that the datasets are just links to copyrighted material, not actual copies of images. That’s true, it’s just scouring the web for images and alt text, but the companies using that data are massively analyzing those images for their own purposes, which are often commercial and valuable. millions of dollars in funding (opens in new tab) to the company.
Getty Images also began legal proceedings this week against stable diffusion (opens in new tab) for the use of copyrighted material in its datasets, which originate from LAION.
But as the class action points out, it’s not just about how these AIs are trained.
“DreamStudio and other image generation products charge users per ‘generation’, but none of that revenue goes to the artists who created the works on which those generations are based. Even if the prompt ‘Dragon in the style of Artist Doe’s, Doe isn’t compensated and they aren’t even consulted or told before these products are built on their job.”
Shutterstock recently announced that it is a AI art tool in collaboration with OpenAI (opens in new tab), the makers of DALL-E and ChatGPT. The tool will allow the site’s users to create AI images, but Shutterstock has said it will also pay artists for their contribution to both training and image creation. Basically, if an artist’s artwork is used in any way by the AI tool, they will receive compensation.
OpenAI’s CEO, who has kept quiet about the datasets the company uses, noted at the time that “the data we licensed from Shutterstock was critical to DALL-E’s training.”
So we can all agree that human-made art is crucial for AI image generators. So how come artists don’t feel fairly compensated for their work?
You can choose not to include your images in some datasets, but that doesn’t undo any training that has already taken place, nor does it mean you’re safe from future scraping. You’d think the obvious answer would be that users have to opt in to be included in a dataset, but that hasn’t been the case with the burgeoning datasets or AI tools out there today. Stability AI intends to make this possible for users opt out with the next version (opens in new tab) of the AI tool.
Your next device
This will be an interesting case to watch, especially as any legal action could set the precedent for further lawsuits related to AI. That said, Ars notes that some parts of the class action complaint may not accurately represent AI technology, and that this case could eventually lose traction over some of these inaccuracies.
Nevertheless, we are only at the beginning of the process and legislative process for the use of AI. Technology is accelerating day by day, finding new applications in all kinds of technology, from Adobe applications to recommendation engines. I wonder if legislation will ever be able to keep up with the pace of AI development.
If you’d like to read the full class action complaint for yourself, you can here (opens in new tab) [PDF warning].