On Tuesday, MEPs adopted a series of recommendations to protect copyrighted creative work from use by artificial intelligence (AI), by 460 votes to 71, and with 88 abstentions. They believe that EU copyright law should apply to all systems of generative artificial intelligence (genAI) on the EU market, regardless of the place of training.
Remuneration and transparency
MEPs insist that use of copyrighted material by genAI must be fairly remunerated in order to protect the EU’s creative sector, which generates 6.9% of the EU’s gross domestic product. They also want the Commission to examine how remuneration for past use can be ensured, but not through a global licence for providers to train their genAI systems in exchange for a flat-rate payment.
MEPs also stress the importance of full transparency for the use of protected content by genAI. They want AI providers and deployers to provide an itemised list of all copyrighted works used to train AI and detailed records of crawling activities for inference and retrieval-augmented generation. Lack of these could be perceived as copyright infringement, triggering legal consequences for AI providers and deployers. If such a court case is then decided in favour of the rightsholder, AI providers or deployers will have to bear all legal costs and related expenses.
Licensing market and opt-out from training
MEPs want the Commission to create a new licensing market for copyrighted material, including voluntary collective licensing agreements per sector, which would include individual creators and small and medium-sized enterprises. They want to make sure rightsholders can exclude their work from being used in AI training and they suggest that the European Union Intellectual Property Office (EUIPO) could manage such an opt-out list.
Protection for news media sector
MEPs urge the Commission to protect the press and news media sector whose work is regularly exploited by AI systems. News media outlets whose traffic and revenues are diverted by AI systems should be fully compensated and they should also have the right to refuse use of their content for training AI systems. MEPs insist that the aggregation of news content must ensure media pluralism and diversity of information, avoiding the selective processing of information or self-preferencing practices by gatekeepers benefiting their AI services.
Content created by AI and individual protection
Content fully generated by AI should not be protected by copyright, MEPs say. They also want to make sure individuals are protected from dissemination of manipulated and AI-generated content and stress the obligation of digital service providers to act against such illegal use.
Quote
Rapporteur Axel Voss (EPP, DE) said after vote: “We need clear rules for the use of copyright-protected content for AI training. Legal certainty would let AI developers know which content can be used and how licences can be obtained. On the other hand, rightsholders would be protected against unauthorised use of their content and receive remuneration. If we want to promote and develop AI in Europe while also protecting our creators, then these provisions are absolutely indispensable.”
Discussion about this post