California judge dismisses parts of lawsuit targeting art-scraping AI

Originally published at: California judge dismisses parts of lawsuit targeting art-scraping AI | Boing Boing

stop motion middle finger GIF by Lawrence Becker


Some of the parts that were dismissed were dismissed for the technical reason that many of the plaintiffs’ images had not been registered with the Copyright Office, not that there was a problem with the underlying legal argument. While copyright in the US is automatic, and registration is not required in order to have a copyright, registration is required in order to recover damages.

I want the artists to get paid, and I think the AI companies did what they were accused of, but this lawsuit it gave me the chills because it seems to imply that learning is copyright infringement and that sounds like a bad idea to have lying around.

The plaintiffs are not arguing that learning is infringement. They are arguing that their works were copied without their permission, which is the very definition of copyright infringement. And, in part, Stability seems to have admitted they did that. From the opinion:

Stability is alleged to have “downloaded of otherwise acquired copies of billions of
copyrighted images without permission to create Stable Diffusion,” known as “training images,”. Over five billion images were scraped (and thereby copied) from the internet for training purposes for Stable Diffusion through the services of an organization (LAION, LargeScale Artificial Intelligence Open Network) paid by Stability. Stability’s founder and CEO “publicly acknowledged the importance of using licensed training images, saying that future versions of Stable Diffusion would be based on ‘fully licensed’ training images. But for the current version, he took no steps to obtain or negotiate suitable licenses.”

It’s important to note that for the images that were registered with the Copyright Office, the lawsuit is being allowed to proceed.


Confused Rooster Teeth GIF by Achievement Hunter


Yep, I know. But that’s the law. You don’t have to register a copyright in order to have a valid copyright, but you do have to register it if you want to recover damages for infringement.


I wonder if the plaintiffs had anticipated the dismissal. Seems like IP law 101 that the copyright has to be registered to claim damages. Doesn’t bode well for this case on the face of it.

This is such a difficult topic. “Downloading” digital copies of art (say, from an artists webpage) to learn about fine art is clearly part of the processing of “learning to art” - if you haven’t seen art, you aren’t likely to have become interested in it, and the idea that acquiring a copy of it (which happens every time your browser shows you an image) being copyright infringement is going a long way to giving all the big players in the art world new weapons to go aver derivative works, which I know isn’t the intent of these actions but I fear is going to be the result. I think ten years from now if lawsuits like this move the needle on copyright aspiring artists are going to be pretty much unable to create art that could be even tangentially similar to a big artists body of work, lest they use these new powers against them.

I really hope I’m wrong, but this is essentially what has happened with music sample analysis.


The truth is that the law has not caught up with the technology. Copyright law largely became necessary in the first place because of the invention of the printing press. It had to be amended with the invention of the phonograph. We have new technology now, and the law needs to adjust. In the meantime, the courts are going to probably have to figure out some compromise. Technology was supposed to free humans from the drudgery of field and factory work so we could go make art and music, and now that same technology is threatening to take that option from us. I don’t have an answer of what to do about this, but we’d better figure it out.


This topic was automatically closed after 5 days. New replies are no longer allowed.