Access to high-quality data for training generative artificial intelligence (AI) in the European Union should be balanced with fair compensation for creators, according to proposals adopted by the Legal Affairs Committee of the European Parliament. The committee voted 17 in favor, 3 against, and 2 abstentions on a set of recommendations aimed at increasing transparency and ensuring remuneration for rightsholders whose copyrighted works are used by generative AI systems.
The proposals call for EU copyright law to apply to all generative AI systems available in the EU market, regardless of where their training occurs. Members of the European Parliament (MEPs) stressed that since generative AI often relies on protected content, there must be full transparency about which copyrighted works are used. They advocate for detailed records from AI providers regarding crawling activities and lists of all copyrighted materials utilized during training. Non-compliance with these requirements could constitute copyright infringement, potentially leading to legal consequences for AI providers.
The committee also highlighted the need for fair payment when copyrighted content is used by AI technologies, aiming to support the sustainability of Europe’s creative and cultural sectors as they adapt to advances in AI. MEPs have asked the European Commission to consider whether such payments should cover past uses as well but rejected a flat-rate global license model that would allow providers broad access to train their systems.
Addressing concerns about media pluralism, MEPs urged both the Commission and member states to safeguard news organizations from potential negative impacts caused by selective news aggregation by AI systems, which can divert web traffic and advertising revenue away from original sources. The report emphasizes that news media organizations should retain control over how their content is used in AI training processes and have the option to refuse permission altogether. It also calls on authorities to ensure appropriate compensation for this use.
MEPs stated that content created entirely by artificial intelligence should not receive copyright protection. They recommended measures designed to protect individuals from manipulated or misleading content generated by AI and called for digital service providers to take action against illegal uses.
The proposals include suggestions for new licensing rules tailored specifically for generative AI applications. The committee called on the Commission to facilitate voluntary collective licensing agreements within different sectors—agreements accessible even to individual creators and small businesses—and explore mechanisms allowing rightsholders to prevent their work from being used in general-purpose AI systems.
Following the vote, rapporteur Axel Voss (EPP, DE) said: “Generative AI must not operate outside the rule of law. If copyrighted works are used to train AI systems, creators are entitled to transparency, legal certainty, and fair compensation. Innovation cannot come at the expense of copyright, both can and must coexist. Clear and enforceable rules are key to ensuring Europe’s technological sovereignty. Our aim is to foster innovation while safeguarding the core principles of intellectual property.”
The own-initiative report will next be presented before a plenary session of Parliament in March.
The report addresses several unresolved legal issues related to how generative artificial intelligence interacts with copyright protections—including questions around transparency requirements, obtaining consent from rightsholders, and establishing fair payment practices when protected works contribute directly or indirectly to outputs produced by advanced machine learning models.
