CJEU Grand Chamber rules on music sampling and pastiche - First CJEU hearing on generative AI and copyright: Like Company v Google

CJEU Grand Chamber defines the pastiche exception in EU copyright law: Kraftwerk v Pelham 

On 14 April 2026, the Grand Chamber of the Court of Justice of the European Union delivered its judgment in case C-590/23. Commonly referred to as 'Pelham II', this case provided a binding interpretation of the pastiche exception under Article 5(3)(k) of Directive 2001/29/EC on the harmonisation of certain aspects of copyright and related rights in the information society (the 'InfoSoc Directive') for the first time. 

The dispute dates back to 1997, when Moses Pelham produced the track 'Nur mir' for the German rapper Sabrina Setlur. The recording incorporated a loop of approximately two seconds taken from 'Metall auf Metall', a phonogram recorded by the German electronic music group Kraftwerk in 1977. This sample was used without authorisation and looped continuously throughout the track. Kraftwerk's members first brought proceedings before the Hamburg courts in the late 1990s. The case was referred to the CJEU for the first time in 2019 (Case C-476/17, Pelham I), when the Court ruled that phonogram producers hold a neighbouring right in their recordings, and that even an extremely short sample may infringe this right, unless a specific exception applies. 

In Pelham II, the Court was asked to rule on whether the use of the sample in question could be considered a pastiche under Article 5(3)(k) of the InfoSoc Directive. The Grand Chamber ruled that 'pastiche' is an autonomous concept in EU copyright law. It defined it as a creation that evokes one or more existing works while remaining recognisably different from them. This exception covers various forms of creative engagement with source material, including stylistic imitations, tributes and works that engage with the original critically or humorously. The defining requirement is that the resulting work must establish an objectively recognisable artistic or creative dialogue with the source material. 

The Court clarified that, when invoking the pastiche exception, a creator need not prove their subjective intention to produce a pastiche. What matters is whether the resulting creation itself objectively demonstrates a creative dialogue with the source work, without requiring the public to look beyond it. The Court also set clear boundaries, stating that the exception does not grant a general licence for artistic appropriation. A reproduction that merely repeats an element of an existing recording without any evident creative interaction with its source does not fulfil the criteria. The case will now return to the German courts, who will determine whether 'Nur mir' qualifies as a pastiche of 'Metall auf Metall' under this test. 

The ruling provides the first clear EU framework for invoking the pastiche exception for businesses operating in the music, audiovisual, advertising and entertainment sectors. Producers, sound designers and content creators who incorporate elements of existing recordings into their work should consider whether the resulting piece establishes an identifiable creative dialogue with the original. Where this is not the case, a licence or authorisation from the rights holder will be required. 

First CJEU hearing on generative AI and copyright: Like Company v Google 

On 10 March 2026, the Court of Justice of the European Union held its first oral hearing on generative artificial intelligence and copyright in case C-250/25, Like Company v Google Ireland Limited. Referred by a Slovak court, the case examines the copyright implications of AI-powered chatbots that display and are trained on copyright-protected press publications. 

Like Company, a Slovak press publisher, alleges that, between 13 June 2023 and 7 February 2024, Google’s AI chatbot, Gemini, systematically extracted and displayed significant sections of its protected press publications in response to user prompts, without authorisation or compensation. The publisher also claims that these publications were used to train the underlying large language model (LLM) without permission.. 

The referring court submitted three questions to the CJEU. The first concerns whether a generative AI chatbot displaying content partially identical to a protected press publication performs an act of communication to the public, as defined in Articles 3 and 15 of the InfoSoc and DSM Directives, respectively. The second question is whether training an LLM on copyright-protected material constitutes an act of reproduction under Article 2 of the InfoSoc Directive. The third question is whether either of these acts is covered by the text and data mining exception in Article 4 of the DSM Directive. This exception permits the use of copyright-protected material for text and data mining purposes, provided that rights holders have the right to opt out. 

The six-hour hearing examined each of these questions in depth. The Advocate General is expected to deliver an opinion on 3 September 2026, with the Court’s judgment to follow at a later date. No ruling has been issued at this stage. 

This case is directly relevant to press publishers, news agencies, online media outlets and any business that produces digital content. The questions before the court will determine whether AI developers operating in the EU must obtain licences before training their models on protected press content and whether the current opt-out mechanism under Article 4 of the DSM Directive provides adequate practical protection. To protect their content from being used in AI training, businesses should apply machine-readable opt-outs to their published material and review the terms of any existing agreements with technology providers.