Getty Images filed a lawsuit against Stability AI in the United States District Court for the District of Delaware on February 3, 2023. The case centers on allegations that Stability AI, the developer of the Stable Diffusion model, copied millions of copyrighted images from Getty's database without permission to train its Generative AI system. The lawsuit seeks statutory damages, injunctive relief, and the destruction of the allegedly infringing training data and models.
Getty Images, a major stock photo agency, claims that Stability AI scraped its website and used its watermarked images to train Stable Diffusion, which can generate new images on demand. The complaint asserts that the AI model reproduces Getty's copyrighted content in ways that violate US copyright law, including direct and contributory infringement. Stability AI has argued that its use of publicly available images constitutes fair use, a defense that has been central to other AI copyright disputes.
Background and Parties
Getty Images, headquartered in Seattle, Washington, is a leading provider of stock photographs, editorial images, and video footage. The company licenses its content to media outlets, advertisers, and creative professionals. Stability AI, founded in 2020 and based in London, is an AI research company known for releasing Stable Diffusion, an open-source text-to-image model that gained widespread attention in 2022. The model was trained on a large dataset called LAION-5B, which contained billions of image-text pairs scraped from the internet, including images from Getty's collections.
The lawsuit was filed in the US District Court for the District of Delaware, a common venue for corporate litigation. The case was assigned to Judge Richard G. Andrews. Getty Images also filed a related lawsuit in the High Court of Justice in London in January 2023, alleging copyright infringement under UK law.
Legal Claims
Getty's complaint includes several causes of action under US copyright law. The primary claims are direct copyright infringement, contributory infringement, and vicarious infringement. Getty alleges that Stability AI directly copied its images during the training process, and that the resulting model enables users to generate images that infringe on Getty's copyrights. The company also claims that Stability AI removed or altered copyright management information, such as watermarks, in violation of the Digital Millennium Copyright Act (DMCA).
Getty seeks statutory damages of up to $150,000 per infringed work, which could amount to billions of dollars given the scale of the alleged copying. The company also requests an injunction to prevent Stability AI from using its images in future training and to require the destruction of the trained model and its derivatives.
Stability AI's Defense
Stability AI has responded by asserting that its use of images for training constitutes fair use under US law. The company argues that the images are used for transformative purposes, as the model does not store or reproduce the original images but rather learns statistical patterns. Stability AI also contends that the training data is publicly available and that the company did not directly copy any specific image. The defense relies on the fair use factors: the purpose and character of the use, the nature of the copyrighted work, the amount and substantiality of the portion used, and the effect on the potential market.
Legal experts have noted that the outcome of this case could set a precedent for the use of copyrighted material in Machine learning training. The case is one of several high-profile lawsuits against AI companies, including actions by authors and artists against OpenAI and other developers.
Procedural History
After the complaint was filed, the court granted a motion to stay discovery pending a decision on the motion to dismiss. Stability AI filed a motion to dismiss in May 2023, arguing that the court lacked personal jurisdiction and that the claims were insufficiently pleaded. In July 2023, Judge Andrews denied the motion in part, allowing the case to proceed on most claims. The court found that Getty had plausibly alleged that Stability AI had purposefully directed its activities at the United States, given that the model was available to US users and that Stability AI had US-based investors.
In early 2024, the court ordered the parties to engage in mediation, but no settlement was reached. The case is currently in the discovery phase, with both sides exchanging evidence and expert reports. A trial date has not yet been set, but observers expect the case to be decided in 2025 or later.
Related Litigation and Context
The Getty lawsuit is part of a broader wave of legal challenges to Generative AI systems. In January 2023, three artists filed a class-action lawsuit against Stability AI, Midjourney, and DeviantArt, alleging similar copyright violations. That case, Andersen v. Stability AI, was partially dismissed in October 2023, but the court allowed the artists to amend their complaint. In December 2023, The New York Times sued OpenAI and Microsoft for copyright infringement over the use of its articles in training Large language models. These cases are being watched closely because they could define the boundaries of fair use for AI training.
Getty's case is notable because the company has a large library of licensed images, and it has taken a proactive stance against AI companies. In 2022, Getty banned the upload of AI-generated images to its platform, and it has partnered with other stock agencies to create licensing models for AI training data.
Implications for AI Industry
The outcome of Getty v. Stability AI could have significant implications for the Artificial intelligence industry. If the court rules that training on copyrighted images without permission is not fair use, AI developers may need to obtain licenses for training data, which could increase costs and slow innovation. Conversely, a ruling in favor of Stability AI could reinforce the practice of scraping publicly available data for training, which is common in the field.
Some companies have already begun to license content for AI training. For example, OpenAI has signed deals with publishers like Axel Springer and the Associated Press. Getty Images itself has launched its own generative AI tool, trained on its licensed content, to offer customers a legally safe alternative. The case may also influence the development of Data Augmentation techniques and the use of synthetic data to reduce reliance on copyrighted material.
Current Status and Future Outlook
As of early 2025, the case remains pending in the Delaware district court. The parties have exchanged initial disclosures and are conducting fact discovery. The court has not yet ruled on the fair use defense, which is likely to be decided on summary judgment or at trial. Legal analysts expect the case to be appealed regardless of the outcome, meaning a final resolution could take years.
The case is also being watched internationally, as similar disputes have arisen in the UK and Europe. In the UK, Getty's parallel lawsuit is still in early stages. The European Union's AI Act, which was adopted in 2024, includes transparency requirements for training data, but it does not resolve copyright questions.
In the meantime, the Machine learning community continues to debate the ethics of training on copyrighted data. Some researchers advocate for open datasets and permissive licenses, while others call for stricter regulation. The Getty case is likely to be a landmark in this ongoing conversation, shaping how AI companies approach data acquisition and copyright compliance.
Conclusion
Getty v. Stability AI (US) is a pivotal copyright lawsuit that tests the legality of using copyrighted images to train AI models. The case raises fundamental questions about the nature of Deep learning and the rights of content creators. Its outcome will influence not only the parties involved but also the broader ecosystem of Generative AI and the future of creative industries. As the litigation progresses, stakeholders across technology, law, and the arts will be watching closely for a decision that could redefine the boundaries of fair use in the age of artificial intelligence.