Getty Images Sues Stable Diffusion Creator Stability AI Over Alleged Copyright Infringement
Stock image service Getty Images is taking legal action against Stable Diffusion, one of the most popular text-to-image tools, over allegations of copyright infringement. The lawsuit has been filed in the High Court of Justice in London and accuses stability AI of unlawfully copying and processing millions of images protected by copyright and the associated metadata owned or represented by Getty Images, without obtaining a license.
Stable Diffusion, which runs on local computers, has gained popularity due to its improved performance on M-powered Macs, which can now generate images in under 18 seconds. However, the increased use of text-to-image generators like Stable Diffusion has raised concerns about potential copyright infringement, as human artists have complained about their works being used without permission or compensation.
In an independent analysis of the 2.3 billion images used to train Stable Diffusion, conducted by Andy Baio and Simon Willison, the researchers found that the model was trained using images from the nonprofit Common Crawl, which scrapes billions of webpages monthly. The analysis revealed that the majority of images came from stock image sites, with 123RF being the biggest contributor, followed by Adobe Stock, PhotoShelter, Dreamstime, iStockPhoto, Depositphotos, Unsplash, Getty Images, VectorStock, and Shutterstock.
The use of copyrighted content for AI training has become a growing concern, with platforms such as Pinterest, WordPress, Blogspot, Flickr, DeviantArt, and Tumblr also found to be sources of images that were scraped for training purposes. The lawsuit against stability AI is just the beginning, and we will likely see a growing number of related lawsuits in the coming months and years unless a balance is found between enabling AI training and respecting the rights of human creators.
To address these concerns, Shutterstock has announced the expansion of its partnership with OpenAI, creator of DALL-E, to create frameworks that will compensate artists when their intellectual property is used and when their works have contributed to the development of AI models.
In conclusion, Getty Images’ lawsuit against stability AI highlights the growing concerns about the use of copyrighted content for AI training and the need to find a balance between enabling AI training and protecting the rights of human creators. The partnership between Shutterstock and OpenAI sets a positive example, and it will be interesting to see if other companies follow suit to ensure that artists are compensated when their works are used in AI training.
