AI that generates images as if they were drawn by human artists is criticized as `` infringing the rights of artists ''
Stability AI , a startup based in London, England and Los Altos , USA, will launch `` Stable Diffusion '', an image generation AI that can create high-quality images that look like they were drawn by a human artist, for researchers. Released. Stable Diffusion is scheduled to be released to the public soon, but human artists have criticized it as 'infringing on the rights of artists.'
On August 10, 2022,
Stable Diffusion launch announcement — Stability.Ai
Stable Diffusion draws controversy on Twitter from artists who say the AI infringes on copyrights
Stable Diffusion is an image-generating AI jointly developed by researchers from Stability AI and Heidelberg University, among others, and trained on LAION Aesthetics , a dataset focused on 'beauty'. “Stable Diffusion is a text-to-image generation model that enables billions of people to create amazing art within seconds,” said Emad Mostaque, co-founder of Stability AI. It is a breakthrough in speed and quality that can be executed on the GPU,' and plans to release it as open source in the future.
The image actually generated by Stable Diffusion is below.
RJ Palmer , who also worked on the creature designs for the movie Detective Pikachu , said, 'The new AI image generator appears to be able to create art that looks 100% human-made. As an artist, I I am very worried, ”he criticized Stable Diffusion for violating the rights of artists.
However, concept artist
A new AI image generator appears to be capable of making art that looks 100% human made. As an artist I am extremely concerned.pic.twitter.com/JUSW0x8Woa—RJ Palmer (@arvalis) August 14, 2022
Palmer's problem is that Stable Diffusion is 'apparently trained by active artists.' If you actually look at the image made by Stable Diffusion exemplified by Mr. Palmer, you can see something like a square frame and broken letters in the lower right. This is a trace that AI tried to reconstruct the 'artist's logo' included in the trained image. 'This AI wants our work and is actively anti-artist.' said.
What makes this AI different is that it's explicitly trained on current working artists. You can see below that the AI generated image(left) even tried to recreate the artist's logo of the artist it ripped off.—RJ Palmer (@arvalis) August 14, 2022
This thing wants our jobs, its actively anti-artist. pic.twitter.com/4zXDeaIUzw
According to Mr. Palmer, hundreds of real artists were listed as 'styles that can be imitated with Stable Diffusion'.
Here is a collection of relatively modern or currently working artists that they advertise as styles to steal on their site(there are hundreds of others).—RJ Palmer (@arvalis) August 14, 2022
In addition, TechCrunch, a technology media, warns that ``Stable Diffusion can also generate fake images and sexual images using the faces of celebrities.''
This startup is setting a DALL-E 2-like AI free, consequences be damned | TechCrunch
This is the `` image of British Prime Minister Boris Johnson with a weapon '' actually created by TechCrunch with Stable Diffusion. The hosted version of Stable Diffusion available through the Discord server seems to prohibit the generation of sexual and racist images, but there is no policy for images containing celebrities.
@StableDiffusion ) is frozen.
TechCrunch said Stable Diffusion's Discord server allowed other image generation services, including images of the war in Ukraine, female nudes, an imaginary Chinese invasion of Taiwan, and religiously controversial images like the appearance of the Prophet Muhammad. It is pointed out that various contents that are not created are generated. ``Apart from illegal content, filtering is minimal, and it's up to you how you use Stable Diffusion,'' Mostaque said.
At the time of writing the article, Stable Diffusion's official Twitter account (