Greg Rutkowski, a digital artist known for his surreal style, opposes AI art but his name and style have been frequently used by AI art generators without his consent. In response, Stable Diffusion removed his work from their dataset in version 2.0. However, the community has now created a tool to emulate Rutkowski’s style against his wishes using a LoRA model. While some argue this is unethical, others justify it since Rutkowski’s art has already been widely used in Stable Diffusion 1.5. The debate highlights the blurry line between innovation and infringement in the emerging field of AI art.
If an image is represented as a network of weighted values describing subtle patterns in the image rather than a traditional grid of pixel color values, is that copy of the image still subject to copyright law?
How much would you have to change before it isn’t? Or if you merged it with another representation, would that change your rights to that image?
It doesn’t matter how you recreate an image, if you recreate someone else’s work that is a violation of copyright.
Stealing someone’s style is a different matter.
Only if the work is copyrighted, and your copy does not constitute fair use…
I could create a faithful reproduction of the Mona Lisa (or… I mean, someone could, I sure couldn’t), and it’s not violating copyright, because the Mona Lisa is not copyrighted.
You could, but Stable Diffusion couldn’t. All it can do is output what it’s been fed. It doesn’t know composition, or colour theory. It doesn’t understand that something is a human, or a fabric, or how materials work, it just reproduces variations of what it’s been fed. Calling it “intelligence” is disingenuous: it doesn’t “know” anything, it just reproduces what’s built into it’s database, usually without the artist’s permission.