In recent months, the European debate on copyright has undergone a significant evolution, particularly regarding the protection of journalistic content in the digital environment, also in light of the growing impact of generative artificial intelligence systems. The discussion, with its most recent development represented by the so-called Voss Resolution, highlights a central issue: ensuring an effective redistribution of the economic value derived from the digital exploitation of content, while safeguarding the economic sustainability of the press sector and the plurality of information.
Directive (EU) 2019/790 (the Copyright Directive) recognized, under Article 15, a related right in favor of publishers, granting them protection over the reproduction and making available to the public of press publications by information society service providers, including automated news aggregators. Conceived in a technological context preceding the widespread diffusion of generative artificial intelligence, these provisions now reveal significant limitations in their application: AI models are capable of using protected content both in the training phase and in the generation of informational outputs, with substitutive effects in relation to direct consultation of sources and potential economic repercussions for publishers and journalists.
The European regulatory framework is complemented by Regulation (EU) 2024/1689, known as the AI Act, which introduces obligations of transparency, traceability, and accountability for providers of AI systems. In particular, the Regulation requires operators of large-scale artificial intelligence models to adopt measures for the documentation, monitoring, and auditing of the data used for training, including sources protected by rights. The provisions contained in the AI Act constitute, for users and, first and foremost, for right holders, initial instruments identified by the legislature to exercise effective control over protected content and to safeguard their intellectual property rights.
Within this context falls the resolution adopted on 10 March 2026 by the European Parliament on copyright and generative artificial intelligence, also known as the Voss Resolution.
The Parliament specifically addresses the impact of generative artificial intelligence on the information sector, recalling the principles enshrined in the Charter of Fundamental Rights, international and EU legislation on copyright and related rights, as well as the need to safeguard the technological and cultural sovereignty of the European Union.
Although it is a non-binding act, the Resolution underscores the ambiguities in the application of the Copyright Directive within such a significantly evolved technological context and the need for swift regulatory clarification, highlighting how the use of content by generative AI systems may have a substitutive effect with respect to the consultation of original sources and may reduce revenues for publishers and journalists. In response to such risks, the Resolution proposes, inter alia, to strengthen transparent the use of protected content for the training of AI models, to promote voluntary collective licensing models in order to ensure fair remuneration throughout the entire creative value chain, and to guarantee full control by right holders over the use of content also beyond the training phase. Attention is devoted to the press considering its fundamental role in safeguarding democracy and democratic structures within the EU.
The Parliament observes that AI systems aggregate and present informational content in an increasingly pervasive manner, with effects on traffic, revenues, and pluralism, recognizing, as a preliminary matter, the need to ensure that GenAI models and systems do not carry out selective processing that favors certain publications over others, thereby preserving the plurality and impartiality of information.
For this reason, the Parliament calls on the Commission and the Member States to protect the information sector by providing for appropriate compensation where providers of GenAI systems divert readership and revenues from press publishers. With regard to copyright protection of journalistic works, paragraphs 5 and 6 of the Resolution reaffirm that right holders, particularly those in the press and media sector – and in particular press publishers, journalists, and editors – must have full control over the digital use of their content by AI systems and models for training purposes, emphasizing that such control should be based on a robust and effective possibility to exclude such use by AI systems and models. The Resolution further suggests that the Commission examine how ancillary rights for press publishers, journalists, and editors, and other related rights, could be extended to additional uses of protected content by artificial intelligence systems and models, such as inference, retrieval-augmented generation, and the production of competing content, ensuring that any exploitation occurs only with the prior consent of right holders and, where appropriate, subject to adequate remuneration, with express reference to the possible collective management of such rights.
The position of the European Parliament thus marks an evolution compared to the original approach of the 2019 Directive: it is no longer merely a matter of strengthening the negotiating position of publishers vis-à-vis platforms, but of ensuring that the economic value generated by the use of content is redistributed among all stakeholders involved in the creation of such content, guaranteeing transparency, effective opt-out mechanisms, and proportionate remuneration. In this way, there is a gradual shift from a model centered on publishers’ related rights to a broader approach aimed at protecting the entire information value chain and all its actors, as well as safeguarding the democratic and cultural values of the European Union.