Getty Images recently announced that it has launched legal proceedings in the High Court in London against Stability AI. Getty claims that Stability AI infringed IP rights including copyright in content that it owns or where it represents the rightsholder:
“It is Getty Images’ position that Stability AI unlawfully copied and processed millions of images protected by copyright and the associated metadata owned or represented by Getty Images absent a license to benefit Stability AI’s commercial interests and to the detriment of the content creators.”
Getty added that it licenses technology developers for purposes related to training AI systems, but that Stability AI did not seek any such licence and instead “chose to ignore viable licensing options and long‑standing legal protections in pursuit of their stand‑alone commercial interests”.
In response a Stability AI spokesman told the Daily Mail: “Please know that we take these matters seriously. It is unusual that we have been informed about this intended legal action via the press. We are still awaiting the service of any documents. Should we receive them, we will comment appropriately.”
Copyright and AI
While we don’t know the details of the complaint yet, the dispute could provide some answers to questions about copyright protection in cases where AI programs create new artwork based on mining millions of images.
In December 2022, Stability AI released Stable Diffusion v2.1, which uses deep learning to enable users to generate detailed images from text.
In its FAQ, Stability.ai says: “The underlying dataset for Stable Diffusion was the 2b English language label subset of LAION 5b, a general crawl of the internet created by the German charity LAION. The CompVis team at the University of Heidelberg trained the model in compliance with German law.”
It adds that there was no opt-in or opt-out for the model data: “It is intended to be a general representation of the language-image connection of the Internet.” However an opt-in and opt-out system is being built for future models.
Stability.ai also says: “The area of AI-generated images and copyright is complex and will vary from jurisdiction to jurisdiction.”
US and EU approaches
Three artists (Sarah Andersen, Kelly McKernan and Karal Ortiz) recently sued Stability AI and other companies for copyright infringement in the United States. This has led to speculation about whether the use of copyrighted material in training AI is protected by the fair use doctrine.
In the EU, Article 4 of the 2019 Copyright in the Single Market Directive states that member states shall provide an exception or limitation to copyright protection “for reproductions and extractions of lawfully accessible works and other subject matter for the purposes of text and data mining”.
This exception/limitation shall apply “on condition that the use of works and other subject matter referred to in that paragraph has not been expressly reserved by their rightholders in an appropriate manner, such as machine-readable means in the case of content made publicly available online”.
This is a very interesting and uncertain legal area, and it will be useful to have guidance from the courts in due course.
To find out more about the issues raised in this blog contact Rosie Burbidge, Intellectual Property Partner at Gunnercooke LLP in London - rosie.burbidge@gunnercooke.com