EU Parliament backs copyright protections for AI training, dividing industry
Tech groups warn of “compliance tax” whilst creative sector celebrates vote demanding transparency and fair remuneration as EU Parliament backs copyright protections for AI training, dividing industry.
Text: Martti Asikainen, 24.3.2026 | Photo by Adobe Stock Photos
The European Parliament has adopted sweeping recommendations to protect copyrighted works from use in artificial intelligence systems, voting by a substantial majority to require AI developers to disclose all training data and compensate creators.
The non-binding resolution, approved 460 votes to 71, calls for creation of a European register where artists can opt out of having their work used for AI training and demands itemised disclosure of every copyrighted work used to build AI models.
The measures have set up a potential clash between creative industries, which welcomed the strong political backing, and technology companies warning the proposals could hamper European competitiveness in the global AI race.
MEPs voted on 10 March to approve measures requiring AI developers to ensure creators receive fair compensation. The resolution aims to address mounting concerns from artists, musicians, and publishers that their work is being exploited without permission or payment.
“We need clear rules for the use of copyright-protected content for AI training,” said Axel Voss, the German MEP who drafted the report. “Legal certainty would let AI developers know which content can be used and how licences can be obtained.”
Transparency requirements
Under the Parliament’s proposals, AI companies would be required to provide itemised lists of every copyrighted work used to train their models, including detailed records of web crawling activities. Failure to comply could be interpreted as copyright infringement, exposing companies to legal action and forcing them to bear all associated costs if they lose in court.
The recommendations also call for creation of a European register, managed by the EU Intellectual Property Office, where rights holders could formally opt out of having their works used for AI training through machine-readable signals.
Parliament rejected the notion of global flat-rate licensing schemes, which some technology companies had proposed as a simpler alternative. Instead, MEPs want sector-specific voluntary collective licensing agreements that would include individual creators and small businesses.
The resolution pays particular attention to press and news outlets, whose content AI systems frequently aggregate and summarise. Parliament argues that media organisations whose traffic and revenues are diverted by AI should receive full compensation and retain the right to refuse use of their content entirely.
The measures also stipulate that content generated entirely by AI should not receive copyright protection—a provision aimed at preserving the distinction between human and machine creativity.
Industry is divided
The vote has exposed sharp divisions between creative industries and technology sector representatives.
Adriana Moscoso del Prado, general manager of GESAC, the European grouping representing authors and composers, welcomed Parliament’s “firm position” in favour of creators’ rights. “Innovation, fairness and cultural sovereignty must go hand in hand,” she told Euronews Next.
The European Composer & Songwriter Alliance said the resolution marked “an important step towards ensuring consent, transparency, and remuneration of rights holders”, whilst calling on the European Commission to propose additional legal frameworks.
However, technology industry associations have warned the measures could hamper European competitiveness. The Computer & Communications Industry Association described the recommendations as potentially creating “a compliance tax” for European companies.
“Today’s non-binding report sends the wrong signal to innovators and risks holding back Europe’s digital competitiveness on the global stage,” said Boniface de Champris, the association’s AI policy lead.
The group argued that existing EU legislation—including the Copyright Directive and AI Act—already provides adequate protections through text and data mining exceptions, which allow AI training on publicly available material unless rights holders opt out.
Economy at risk
The creative sector accounts for 6.9% of the EU’s gross domestic product, according to Parliament figures, giving the debate significant economic weight beyond questions of artistic rights.
Parliament’s recommendations now go to the European Commission, which is currently reviewing the EU’s copyright framework. The Commission must decide whether to propose binding legislation based on the resolution or rely on enforcement of existing rules.
The debate reflects broader tensions across Europe as policymakers attempt to balance fostering AI innovation with protecting traditional industries and cultural production. Similar disputes are unfolding in the United States and United Kingdom, where creative industries have filed lawsuits against AI companies over unauthorised use of copyrighted material.
Parliament’s Legal Affairs Committee had worked on the report since June 2025, conducting workshops and commissioning research studies on the intersection of generative AI and copyright law. The overwhelming vote margin—with 460 MEPs supporting the resolution—suggests strong political backing for measures to protect creative works, though implementation will depend on the Commission’s response.
