On This Day in AI: January 17, 2023
On January 17, 2023, Getty Images announced from London that it had begun legal proceedings in the High Court of Justice against Stability AI, the company behind Stable Diffusion. Getty said Stability AI had unlawfully copied and processed millions of copyrighted images, captions, and associated metadata from its collections to train an image-generating system. The photo agency argued that it licenses digital material for many kinds of technology development and that the model maker had not sought a license for this use. The announcement opened a major dispute between a global visual archive and one of the most prominent developers in the fast-growing generative-image market.
Getty’s business depends on maintaining records about who created an image, what it depicts, and how a customer may use it. Generative models depend on very large datasets from which they learn patterns linking pictures and language. Those systems made it possible to create a new-looking image from a short prompt, but they also raised questions about how training material had been obtained. Getty pointed to generated pictures that distorted its watermark as evidence that its content had influenced the system. A false or malformed watermark did not by itself prove every legal claim, yet it made the conflict visible to ordinary users and highlighted the value of captions and metadata alongside the pixels.
The January 17 announcement started litigation; it did not settle whether training on copyrighted images was lawful. Getty later filed a related United States case, while Stability AI contested the claims and the courts examined jurisdiction, copying, trademarks, and the technical behavior of the model. The proceedings joined a wider argument over fair use, licensing, competition, and whether model outputs could substitute for the markets of original creators. The case also encouraged publishers and technology companies to consider negotiated data licenses, clearer dataset documentation, opt-out mechanisms, and content provenance. Its legacy lies partly in the identity of the plaintiff: Getty was not an individual artist with limited records, but a sophisticated licensing company with a large catalog and established enforcement practices. The dispute showed that generative AI would have to confront the economics and legal structure of the media industries whose material helped make model training possible, even as courts worked slowly through unfamiliar technical evidence.
The conflict also affected ordinary buyers of synthetic images. Businesses needed confidence that a generated asset would not expose them to unexpected claims or imitate a protected brand. That commercial uncertainty gave developers an incentive to document data rights and offer clearer warranties. The case therefore connected creator compensation with the practical adoption of generative tools: customers, investors, and platforms all benefited when the source and permitted use of training material could be explained rather than assumed.
Getty alleged that millions of licensed images, captions, and metadata had been copied without permission to train Stable Diffusion.
The company emphasized that it already licenses material for technology uses, making consent and commercial terms central to its complaint.
Some generated pictures contained distorted marks resembling the Getty watermark. The artifacts made training-data influence easy to see, even though the legal issues required broader evidence.
A watermark resemblance did not decide whether every training or output act was infringement. Courts still had to examine technical processes and specific legal claims.
The action increased pressure for licensed datasets, provenance records, opt-out tools, and clearer explanations of what material trained commercial systems.
It also showed that established media companies would use copyright and trademark law to defend archives whose value depended on traceable rights and reliable metadata.
Explore more of "On This Day ..."
Discover more events from the same date across news, politics, technology, sports, and other fields. Each link highlights significant moments that shaped history on different fronts.