The mixed ruling was claimed as a victory by both Stability AI and Getty Images, muddying the legal waters around AI and intellectual property.
The United Kingdom’s High Court of Justice issued a mixed ruling on Tuesday in the Getty Images v. Stability AI intellectual property case filed in 2023, which mostly favored Stability AI but left key questions unanswered about AI’s use of copyrighted material.
Getty owns a library of copyrighted online stock images — which it licenses to users for a fee — and alleged that Stability’s Stable Diffusion AI model, which is trained using online material, infringed upon its trademark and copyrighted material.
Stability’s Stable Diffusion AI model infringed on Getty’s trademark by reproducing its watermark in certain cases. However, the findings were “extremely limited in scope,” Justice Joanna Smith ruled.
Getty failed to show that any UK users used Stable Diffusion to reproduce the watermark, which is required under UK law to prove “primary infringement,” she ruled.
Smith also dismissed the “secondary infringement” allegation because the AI model does not store or reproduce the images, failing to satisfy the requirements for a violation under the UK’s Copyright, Designs and Patents Act (CDPA) of 1988. She wrote:
“Although an ‘article’ may be an intangible object for the purposes of the CDPA, an AI model such as Stable Diffusion, which does not store or reproduce any Copyright Works, and has never done so, is not an ‘infringing copy’ such that there is no infringement under sections 22 and 23 CDPA.”
The ruling leaves the door open for brands to protect their trademarks from AI reproduction, but the technicalities in the case prevent a broad legal precedent from taking effect, leaving key questions about AI training and intellectual property open for debate.
Related: Crypto czar David Sacks argues AI threat is Orwellian, not Terminator
Blockchain and Web3 could provide a critical lifeline to content creators to protect their work
US judge William Orrick issued a similar ruling in October 2023, dismissing most copyright infringement claims against Midjourney AI, DeviantArt and Stability AI.
Orrick said that images generated by AI models do not constitute copyright infringement because they do not bear a resemblance to the artists’ original work on which the models are trained.
The lack of legal protections for content creators and artists has prompted several blockchain and Web3 companies to create data provenance solutions to record ownership and verify sources of information, copyrighted material and other intellectual property.
These include non-fungible tokens (NFTs), which can be used to track the original ownership and assign royalty rights for art, essays, books, musical productions and other creative works.
