Copyright Infringement in AI Models: Statutory interpretation, Memorisation and Metaphors in Getty Images and GEMA

Abstract

Artificial intelligence (AI) is the latest technology to confront the boundaries of copyright regulation. Academics, practitioners, industry, rightsholders and policymakers are grappling with the questions of copyright’s applicability to AI globally. Meanwhile over 85 cases have been filed in the US to litigate these questions, with two decisions made in Europe: one in the UK and the other in Germany. This article explores the scope of copyright infringement in the context of artificial intelligence. It provides analysis of the UK case of Getty Images v Stability AI, which found that AI model Stable Diffusion was not committing secondary copyright infringement. It contrasts this outcome with the decision in GEMA v Open AI from the Munich Court in Germany, which found that Chat GPT was infringing the copyright of music rightsholders. The analysis is undertaken through the lens of statutory interpretation and its impact on the case outcomes, as well as the influence of metaphorical devices when fact-finding in copyright litigation and the narratives around technology, in particular the notion of AI ‘memorisation’ which is a pivotal issue in both matters.