email print share on Facebook share on Twitter share on LinkedIn share on reddit pin on Pinterest

INDUSTRY / MARKET Europe

The Council of the European Union publishes a report on the relationship between generative artificial intelligence and copyright

by 

- Member states have contributed to a policy questionnaire to discuss laws concerning ownership of AI content, highlighting key concerns set to impact audiovisual production and artistic creation

The Council of the European Union publishes a report on the relationship between generative artificial intelligence and copyright
(© European Audiovisual Observatory)

The Hungarian presidency of the Council of the European Union published and distributed a questionnaire to EU member states on 27 June 2024, inviting them to submit proposals on seven different topics in the area of generative artificial intelligence (AI) and copyright by 1 November 2024. As a common outcome, most member states agreed that the current EU legal framework is sufficient to address emerging issues involving AI. However, greater clarity and certainty on the topic of AI would be beneficial in the process. The majority of member states declared that there are no significant court decisions or case law on the topic in their respective countries, except for three cases. The preferred approach would be to harmonise regulations across all EU countries, rather than addressing them solely at the national level, as variations in regulations could prove counterproductive.

(The article continues below - Commercial information)
focusfeatures_conclave_Internal_Cathy

The first question asked member states about the existence of relevant surveys, studies or research conducted on the relationship between generative AI and copyright. The results show that such studies are frequent, with some members providing first-hand examples. Some of the findings in this area are particularly notable. For instance, a study conducted on behalf of a CMO (collective management organisation) representing music creators and rightsholders revealed that 95% of adults believe the state should adopt a proactive approach to the effects of AI on the music industry. Another survey indicates that an impressive 92% of the authors surveyed can recognise AI-generated visuals, texts or music.

In the second set of questions, focusing on AI model training, one member state highlighted that, although the EU was the first to regulate this area, most training occurs outside the EU. This raises concerns about whether rightsholders are adequately informed or have consented to such activities. Another member state and a stakeholder expressed doubts about how EU solutions would apply to AI systems trained outside the EU.

Regarding language diversity, two member states made reference to the ALT EDIC project under the European Commission’s European Digital Decade Programme. The primary aim of this project is to ensure the survival of minority European languages in the AI domain while promoting linguistic and cultural diversity in data training.

Furthermore, member states agreed that purely AI-generated works cannot be protected by copyright, in line with the Berne Convention. As such, only a natural person can be recognised as an author. Nevertheless, the majority of contributors supported granting copyright protection to AI-assisted works, provided the standard requirements for copyright protection are met. This distinction is deemed crucial, with the recommended approach being assessment on a case-by-case basis.

Two distinct perspectives emerged regarding labelling: one group of member states advocated labelling content exclusively generated by AI, while the other emphasised the need to differentiate between AI-generated and AI-assisted content. A significant number of member states and stakeholders stressed that labelling should be adapted to the type of work or medium. However, another sizeable group of member states argued that such differentiation is unnecessary at this stage, primarily owing to technical challenges.

In response to the question of whether a specific liability regime is needed for copyright infringement related to generative AI activities, all members highlighted the importance of enforcing existing rules, rather than introducing a new liability regime. A few member states drew attention to the rise of cloning models and stressed the need to address issues such as deepfakes and the unauthorised use of individuals’ voices, images, names and likenesses in AI processes.

Finally, member states unanimously agreed on the importance of holding an international dialogue on AI and copyright, underscoring the necessity of the EU’s active participation and contribution to this discussion.

The full document is available to read here.

Did you enjoy reading this article? Please subscribe to our newsletter to receive more stories like this directly in your inbox.

See also

Privacy Policy