In the latest episode of Carpmaels in Conversation, Lara Elder and Harry Shaw discuss one of the most closely watched cases in intellectual property law: Getty Images v Stability AI. Using the case as a starting point, they examine a central question for modern copyright law: whether AI models trained on datasets that include copyright works should give rise to liability for infringement.

The episode explains how generative AI models such as Stable Diffusion are trained on vast quantities of data, and explores the key legal issue at the heart of the dispute – whether AI models can give rise to ‘infringing copies’. It also considers what the decision indicates about the limits of existing copyright frameworks in the context of AI training, and why the case is likely to remain significant as the law continues to develop. You can read our article on the case and its implications here.