• The VOIDRUNNER'S CODEX is coming! Explore new worlds, fight oppressive empires, fend off fearsome aliens, and wield deadly psionics with this comprehensive boxed set expansion for 5E and A5E!

AI is stealing writers’ words and jobs…

[Andersen v Stability] Midjourney files motion to dismiss largely on procedural and technical grounds

Midjourney filed a motion to dismiss the claims of Andersen et al in Andersen v Stability (they are one of the defendants).

https://storage.courtlistener.com/recap/gov.uscourts.cand.407208/gov.uscourts.cand.407208.160.0.pdf

Choice words from the motion's introduction:

The FAC adds only filler, not facts to support a viable claim against Midjourney. Rather
than curing the pleading defects noted in the Court’s October 30, 2023 dismissal order (“Order”),
the FAC is larded up with new plaintiffs (many of whom lack a copyright registration for the images
at issue), newly manufactured and misleading exhibits (showing only that plaintiffs lack facts to
substantiate their claims), and new theories of liability that contradict their earlier pleading and are
just as baseless as those already rejected.
One of the larger bases for the motion is that cited instances of supposed infringement fail to denote exactly which works were infringed upon, and examples of infringement do not match any registered works.

In other words, say you painted an apple, a banana, and an orange in Your Style. You only register the work of an apple. In a lawsuit against a genAI company, you point to outputs that look a lot like your banana and your orange. Defense then goes and says "that may well be, but you didn't register banana/orange, and while you did register apple, the fact that banana/orange have a similar style to the registered apple-in-Your-Style work is moot because styles can't be copyrighted.

Similarly: books that were registered but only covered the text, not the images / registrations of collections of works which applies only to the collection not the pre-existing material within the collection / Class claims against 'any defendant' being an inappropriate filing for any one of the specific defendants, alleging a genAI model to be a derivative work when it fails even the substantial similarity test, and so on.

Basically just the latest round between the Plaintiffs and Defendants, but it's interesting to me that - while entirely reasonable and certainly the first sort of thing any lawyer should go for - procedural and technical issues are at the forefront of the motion, rather than a direct defense.

While the motion does have legalese and makes other case references, their argument should be fairly understandable to anyone.

Reminder: This is just a motion by the defendants, not any court opinion or judgment
__



There are also new motions to dismiss from all of them - Stability, Midjourney, DeviantArt, and Runway.

a few snippets from the various motions:

The Stability one has a great line:

"Nearly a year after filing their initial complaint, the only examples of purportedly infringing outputs Plaintiffs include in the FAC are those created by Plaintiffs’ attorneys themselves."

and other motions

"The fact that Plaintiffs were able to generate “Image Variations” from an “Image Variation” service does not prove that the underlying model is a “copy” of those images — any more than a photocopier’s ability to duplicate a photograph proves that the photocopier itself is a “copy” of that photograph. Instead, even assuming the images in Exhibit H are substantially similar to Plaintiffs’ works, Plaintiffs’ creation of those images does not mean that the parties who implement the models Plaintiffs used are somehow directly liable for copyright infringement. If anything, it suggests that Plaintiffs just “reproduced” “copies” of their own works."

"Plaintiffs have been unable to elicit any “stored” copies of their own registered works from Stable Diffusion, despite ample opportunities to try. And that is fatal to their claim. The question is not whether Stable Diffusion “stores” any image. Rather, to support their Model Theory, Plaintiffs needed to plead facts showing, at a minimum, that Stable Diffusion “stores” protected expression from Plaintiffs’ registered works.

Plaintiffs have had ready access to Stable Diffusion 1.5 and to Runway’s platform that they allege uses it. But nowhere do they allege that they, or anyone else, have been able to elicit replicas of their registered works from Stable Diffusion by entering text prompts. Plaintiffs’ silence on this issue speaks volumes, and by itself defeats their Model Theory."
 

log in or register to remove this ad


overgeeked

B/X Known World
People are hilariously un-self-aware. How about you track down and credit the artists whose work this is based on and credit them first?

BE72D00B-6ABB-4289-A3D7-DD07AA62EBEA.jpeg
 





Scribe

Legend
I'm sure this goes against the over theme of this thread but



Meanwhile in India https://pib.gov.in/PressReleasePage.aspx?PRID=2004715 AI works are covered by existing copyright law. Individuals aren't infringing copyright unless they generate infringing works.

Ah good, so I can have ChatGPT rip most of the D&D published body of work for my own RPG and sell that. Fantastic.
 

Ah good, so I can have ChatGPT rip most of the D&D published body of work for my own RPG and sell that. Fantastic.
"Our Plaintiffs' core claim against OpenAI—direct copyright infringement under the U.S. Copyright Act for training language models using Plaintiffs’ copyrighted works—was not dismissed"

Among copyright claims tossed by Martínez-Olguín were accusations of vicarious copyright infringement. Perhaps most significantly, Martínez-Olguín agreed with OpenAI that the authors' allegation that "every" ChatGPT output "is an infringing derivative work” is "insufficient" to allege vicarious infringement, which requires evidence that ChatGPT outputs are "substantially similar" or "similar at all" to authors' books.

"Plaintiffs here have not alleged that the ChatGPT outputs contain direct copies of the copyrighted books," Martínez-Olguín wrote. "Because they fail to allege direct copying, they must show a substantial similarity between the outputs and the copyrighted materials."

Authors also failed to convince Martínez-Olguín that OpenAI violated the Digital Millennium Copyright Act (DMCA) by allegedly removing copyright management information (CMI)—such as author names, titles of works, and terms and conditions for use of the work—from training data.

This claim failed because authors cited "no facts" that OpenAI intentionally removed the CMI or built the training process to omit CMI, Martínez-Olguín wrote. Further, the authors cited examples of ChatGPT referencing their names, which would seem to suggest that some CMI remains in the training data.

Some of the remaining claims were dependent on copyright claims to survive, Martínez-Olguín wrote.

__
When/if it gets to discovery first thing I'd ask for is the full chat log from the session where they used the bot to produce purportedly infringing work.
 


Remove ads

Top