The problem of pleadings in AI copyright litigation: lessons from <i>Getty v Stability</i> (UK) and <i>Doe v GitHub</i> (US)

Abstract This article considers the difficulties facing authors who claim their copyrighted works have been infringed by an artificial intelligence (AI) provider. In particular, it looks at the tension inherent between the extent of factual detail required by local rules of practice for initiating pleadings and the ‘black box’ nature of generative AI tools—notwithstanding mandatory reporting requirements such as those under Article 53(1)(d) of the EU’s Artificial Intelligence Act (Regulation (EU) 2024/1689). Drawing from interim decisions in the UK’s Getty v Stability and the USA’s Doe v GitHub, it will consider whether Rules of Court which commonly require a detailed description of the wrong allegedly committed by the defendant(s) present as a bar to claimants. It will be argued that interlocutory discovery applications, known in some jurisdictions as ‘disclosure’, do not currently provide an adequate pathway for prospective claimants to obtain documentary evidence, which is probative to their intended case. Consequently, a new type of pre-action procedure is needed to rebalance the rights of authors against operators of emerging technologies.