Getty Images, one of the world’s largest image agencies, has sued Stability AI, the company behind Stable Diffusion generative AI. This has been announced by the agency itself through a statement, where it explains that it has filed a lawsuit in the High Court of Justice in London, for copyright infringement.
Generative AIs use thousands of images from the internet to train their algorithm. Getty Images explains that millions of its protected images have been used, without permission. The agency trusts in AI and has in fact worked with generative art systems, but through concrete agreements and respecting intellectual property rights. However, Stability AI did not seek any type of agreement to license the images from Getty Images.
The agency does not believe that the use of these images falls under “fair use”, but rather directly constitutes an infringement of their intellectual property. The step of denouncing Stability AI represents an escalation in the legal battle between the creators and the generators of images with AI.
This is not the first complaint against AI. Just this week we learned that a group of creators, together with a lawyer who is already carrying out a parallel lawsuit against Github Copilot, have denounced both Stability AI and Midjourney.
The position of the creators of Stable Diffusion is that their datasets have been trained “ethically, morally and legally”. Although, they recognize that in the next version of Stable Diffusion they will allow an artist to withdraw their works from the training set. Something that is not in the current version.
The new lawsuits against Stable Diffusion may end up contributing to the generation of a new legal framework. Currently there is not enough jurisprudence regarding copyright and its relationship with AI. While open generative AIs such as Midjourney or Stable Diffusion have been denounced, the same is not the case with OpenAI’s DALL-E, which maintains numerous agreements with image agencies and stock image banks to train its algorithms.
Image | Stable Diffusion via The Verge
