AbstractArtificial intelligence is an emerging technology with an average growth rate of 64.70% in patent filings worldwide and 39.71% in Europe between 2015 and 2019, according to Espacenet queries. This trend has raised several concerns regarding the implications for Intellectual Property rights, including disclosure, which is one of the main components justifying the existence of the patent system. Despite its importance, this requirement has been disputed due to certain tendencies that distort the functions that the act of disclosure fulfils, such as promoting innovation, research or teaching. With artificial intelligence–related patents being granted, the disclosure requirement is once again compromised, and perhaps even more due to the particularities that this branch of technology entails. In this sense, one main concern lies in the question of how artificial intelligence–related patents should be disclosed, as well as the fulfilment with this requirement. To this aim, two case studies are conducted to show the heterogeneity of compliance with this requirement, focusing on patents filed at the European Patent Office. The methodology used in the case studies consists in the analysis of a core-AI patent and an AI application patent which allow to assess the differences in the disclosure of both inventions. The heterogeneity between both case studies highlights the relevance of the topic and allows to propose three recommendations to improve sufficiency of disclosure in AI-related patent claims: metrics and benchmark analysis, standardisation of patent claim construction by introducing a ‘claim chart’ and introducing a peer review patent programme.