Getty Images v Stability AI: High Court ruling poses questions for creative industry
The High Court has largely ruled in favour of Stability AI in the intellectual property dispute brought by Getty Images over the use of stock images to train the Stable Diffusion AI model. Getty claimed that Stability AI infringed by training its AI on millions of images from its library.
The Court found that using copyrighted material to train an AI model, without storing or reproducing the works in the model itself, does not constitute secondary copyright infringement.. Mrs Justice Joanna Smith stated, “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’.”
The Court did, however, agree that AI-generated images containing Getty watermarks were considered infringing, though this applied only to older models and isolated instances.
Speaking with multiple publications, Data Protection, AI and Digital Regulation Partner James Clark commented on the decision.
“At the heart of the Getty Images judgment is the finding that the training of Stable Diffusion’s AI model using copyright work did not result in the production of an infringing copy of that work,” James said.
“At the end of the training process, the AI model did not store any copy of the protected works, and the model itself was not itself an infringing copy of such work. It is this finding that will cause concern for the creative industry whilst giving encouragement to AI developers.
“The judgment usefully highlights the problem that the creative industry has in bringing a successful copyright infringement claim in relation to the training of large language models.
“During the training process, the model is not making a copy of the work used to train it, and it does not reproduce that work when prompted for an output by its user. Rather, the model ‘learns’ from the work, in a similar way to the way that you or I might do so.
“As an expert report quoted in the judgment explains: “Rather than storing their training data, diffusion models learn the statistics of patterns which are associated with certain concepts found in the text labels applied to their training data, i.e. they learn a probability distribution associated with certain concepts.”
The ruling leaves unresolved questions for the creative sector. Getty withdrew part of its claim relating to primary copyright infringement due to jurisdictional limits, and further proceedings continue in the US.
Read the coverage and James’ full comments below:
