EU Copyright Law and Generative AI: A Watershed Moment

Generative artificial intelligence (GenAI) has pushed EU copyright law into a full-blown crisis. AI models are trained on millions of copyrighted works, often without a licence, without payment, and without transparency. On 25 February 2026, the European Parliament's Committee on Legal Affairs published its report on "Copyright and Generative AI – Opportunities and Challenges" (Rapporteur: Axel Voss). The vote result was clear: 17 in favour, 3 against.

The document goes far beyond a political statement. It sketches a fundamental re-orientation of European copyright law — and makes concrete demands of AI providers, rights holders, and the legislature.

The Problem: Existing Law Does Not Fit the Technology

GenAI is a type of AI that, unlike other AI systems designed primarily to classify or predict, generates new content – such as text, images, music, videos and code – on the basis of training using very large datasets. GenAI outputs are created through predictions based on statistical models, generally mimic human creativity, and rely on pre-existing content, which may include copyright-protected materials.

The existing legal framework was not designed for this reality. The current systems for the reservation of rights ('opt-out') to copyright-protected content are often impractical, may not cover all relevant acts of text and data mining, and lack the necessary transparency for effective implementation and enforcement.

The scale of non-compliance is striking. There is evidence of the widespread violation of copyright rules by GenAI providers, including the unauthorised collection of works from the internet, the non-compliance with rights holders' text and data mining rights reservations, the use of pirated sources to obtain works, and the failure to seek licences.

The Parliament draws a stark conclusion: current copyright law is insufficient to address the challenge of licensing copyrighted material for GenAI, and an additional legal framework is needed to clarify licensing rules for GenAI and to address potential infringements.

Territoriality Reimagined: EU Law Follows the Market

One of the most far-reaching proposals concerns the territorial reach of EU copyright law. AI training frequently takes place outside the EU, which has made it difficult for European rights holders to enforce their rights.

The Parliament now calls for a direct solution. To ensure a level playing field between EU- and non-EU-based GenAI services, the principle of territoriality must be adapted to the training of generative AI systems so as to ensure that the use of European content is subject to EU law even when such training takes place outside the EU. GenAI systems that do not comply with these requirements should be barred from operation within the EU.

Rules on copyright and related rights need to apply uniformly to all AI providers deploying products or offering services within the EU, irrespective of their place of establishment, the jurisdiction in which the copyright-relevant acts underpinning the training of those AI models take place, and where the output produced by the AI system is used in the EU.

This is a fundamental shift: the market-of-destination principle would displace the place-of-origin principle. For non-EU AI companies – including those training models in the United States or Asia — the implications are significant.

Transparency as a Core Legal Obligation

Transparency is the centrepiece of the report's enforcement architecture.

Any provider or deployer of GenAI models and systems should ensure transparency concerning all copyright-protected content used to train those models and systems, irrespective of the jurisdiction in which the copyright-relevant acts underpinning the training take place. This transparency must consist of an itemised list identifying each item of copyright-protected content used for training.

The obligation goes beyond training. For purposes such as inferencing and retrieval-augmented generation (RAG), which require continuous and real-time crawling, transparency should include the act of crawling itself, whereby crawlers are required to identify themselves to web operators, and AI companies are required to maintain detailed records of their crawling activities.

The report also addresses the widespread practice of invoking trade secrecy to avoid disclosure. The mere information of the third-party content used by AI and GenAI providers and deployers does not constitute trade secrets under EU law.

Failure to comply triggers serious legal consequences (see below).

The Rebuttable Presumption: A Game-Changer for Enforcement

For rights holders, one of the most practically significant proposals is the introduction of a rebuttable presumption.

The Commission should be called on to propose the establishment of a rebuttable presumption that, for any generative AI model or system placed on the EU market, works and other subject matter protected by copyright or related rights have been used for the purposes of training, inferencing or retrieval-augmented generation, where the transparency obligations set out in the resolution have not been fully complied with.

The cost consequences are equally notable. Where a court finds in favour of a rights holder or the organisations representing them on the basis of either such a presumption or of submitted evidence, all reasonable and proportionate legal costs and other expenses should be borne by the AI provider.

This reversal of the burden of proof — combined with cost liability — could fundamentally alter the litigation landscape in IP disputes involving AI.

EUIPO: A New Central Role

The European Union Intellectual Property Office (EUIPO) is positioned to become a key institution in the new framework.

The Parliament proposes making the EUIPO the trusted intermediary that manages and lists opt-out exclusions, providing AI providers with a comprehensive tool to ensure copyright compliance, while participation in any new mechanism should be simple and cost-efficient and should not invalidate or override previously expressed refusals.

Beyond the opt-out register, the EUIPO would take on a broader role. The Commission should assign to EUIPO responsibility for supporting a sector-based, voluntary licensing process, so as to streamline relations between GenAI providers and rights holders, establishing a workable, innovation-friendly framework that supports the EU's competitiveness.

The EUIPO Copyright Knowledge Centre will play a vital role in guiding the use of copyright in the age of GenAI by raising awareness, promoting legal clarity, and fostering a balanced framework that supports creativity, innovation, cultural preservation, and European competitiveness.

Building a Licensing Market – Including for Past Uses

A functioning licensing market is a central objective. The Commission should facilitate, through consultation with collective management societies, the establishment of voluntary collective licensing agreements per sector as a means to quickly establish a working licensing market that provides fair remuneration for rights holders while enabling AI providers to access high-quality training data.

The question of retroactivity is particularly significant. The Commission should examine whether there is a possible solution for the immediate, fair and proportionate remuneration for past uses of copyright-protected works by providers of general-purpose AI models and systems where a licensing market could not yet be established. The Parliament explicitly rejects a global licence with flat-rate payments as a model.

The Explanatory Statement goes further still. Pending the introduction of an appropriate provision, the Commission should establish an immediate, simple, flat-rate copyright fee of 5 to 7% of global turnover in order to compensate for the added value that these businesses generate using the data of European creatives and to ensure it remains in Europe.

AI-Generated Content: No Copyright Protection Without Human Authorship

The report draws a clear line on the legal status of AI outputs.

Content fully generated by AI that does not meet the established criteria for copyright protection should remain ineligible for copyright protection, and the public domain status of such outputs should be clearly determined.

EU copyright law remains grounded in the principles of human authorship. According to the settled case-law of the Court of Justice of the European Union, the concept of a 'work' entails two cumulative conditions: first, it must be an original subject matter that reflects the author's own intellectual creation, and second, that creation must be expressed in a manner that makes it identifiable with sufficient precision and objectivity.

The Press Sector: A Specific Area of Concern

Rights holders, especially from the press and news media sector, must have full control over the digital use of their content by AI systems and models for training purposes, based on a robust and functioning possibility to exclude such use, underpinned by full transparency and source documentation.

Rights holders must also have full control over the use of their content for purposes beyond AI training, such as inferencing and retrieval-augmented generation. The Commission should explore how ancillary rights for press publishers could be extended to cover those purposes.

The Commission should explore mechanisms ensuring that providers of GenAI models or systems that demonstrably divert traffic and revenue from press and news media outlets compensate such outlets in a fair, proportionate and non-discriminatory manner, paying particular attention to local and regional media.

The Long-Term Vision: A General Copyright Regulation?

The Explanatory Statement raises an ambitious long-term possibility. A 'General Copyright Protection Regulation', akin to the General Data Protection Regulation (GDPR) – or rather the European Union Trade Mark Regulation, could be helpful in creating a permanently workable solution and avoiding the need to protect copyright claims against new technologies every five or six years.

More immediately, copyright law needs to be adapted to technological developments, which will require further European harmonisation of Member States' national copyright laws.

What This Means in Practice

The report is not yet binding law — but it is a powerful political signal, and it points in a clear direction.

• For AI providers: Detailed transparency obligations over training data are coming — with severe consequences for non-compliance, including potential market exclusion from the EU and a reversed burden of proof in litigation.
• For rights holders: The opt-out system is set to become stronger, more standardised, and machine-readable. A central register at the EUIPO will provide greater legal certainty.
• For publishers and media organisations: Control over content is to be extended beyond training to inferencing and RAG. Ancillary rights may be broadened accordingly.
• For all businesses using AI systems: Copyright compliance obligations will apply not just to model providers but also to deployers of AI systems.

The Commission should urgently conduct a thorough assessment of whether the existing EU copyright acquis adequately addresses the legal uncertainty and competitive effects associated with the use of protected works for the training of GenAI systems, independently of and before starting its planned review of the CDSM Directive.

The window for proactive action is open. Businesses across all sectors should use it.