The rapid rise of generative artificial intelligence (GenAI) is forcing a fundamental rethink of copyright law in the European Union. A recent report by the European Parliament’s Committee on Legal Affairs marks what many consider a “watershed moment”—not merely clarifying existing law, but signalling a potential overhaul of how copyright operates in the age of AI.
The Core Tension: AI vs Copyright
Generative AI systems—capable of producing text, images, music, and code—are trained on vast datasets, often scraped from the internet. These datasets frequently include copyrighted works, used without explicit permission, compensation, or even transparency.
This creates a fundamental legal conflict:• Copyright law is built on protecting human creativity and control over works • Generative AI depends on mass ingestion of existing creative content
The result is a structural mismatch. Existing legal frameworks were never designed for machines that learn by absorbing millions of protected works and then generate new content that may compete with the originals.
Why Existing EU Law Falls Short
EU copyright law—particularly the 2019 Copyright Directive—contains text and data mining (TDM) exceptions, allowing certain uses of copyrighted material without permission.
However, these provisions are now under strain:
• They were designed for research and limited commercial uses • They did not anticipate large-scale commercial AI training • Rights holders often lack practical means to opt out or enforce rights
This has led to what policymakers increasingly describe as a legal gap, where AI developers can rely on exceptions that were never intended for their use case.
The European Parliament’s Intervention
In February 2026, the European Parliament’s legal affairs committee adopted a report addressing “Copyright and Generative AI – Opportunities and Challenges.”
The report is significant because it goes beyond commentary—it proposes a reorientation of copyright law, with concrete expectations for: 1. AI Developers• Greater transparency about training data • Compliance with copyright rules across jurisdictions • Potential obligations to license content 2. Rights Holders• Strengthened ability to control or reserve rights • Improved tools to detect and enforce infringement 3. Legislators• A need to update copyright frameworks to address AI specifically • Clarification of how existing exceptions apply (or don’t apply) to GenAI A Shift in Territoriality: “Follow the Market” One of the most important emerging ideas is a shift in how copyright jurisdiction works. Traditionally, copyright law is territorial—linked to where an act (like copying) occurs. But AI complicates this:• Training may occur outside the EU • Outputs are used within the EU market The proposed shift suggests that EU copyright rules may follow the market rather than the location of training. This would mean:• Non-EU AI providers could still be subject to EU copyright law • Non-compliant systems could be excluded from the EU market This represents a major expansion of the EU’s regulatory reach. Transparency as a Central Requirement A key theme is transparency. Currently, creators often cannot determine:• Whether their work was used in training • How it was used • Whether outputs compete with their work The Parliament is pushing for:• Disclosure of training data sources (at least in summary form) • Mechanisms for rights holders to identify and assert claims Transparency is seen as a prerequisite for any meaningful enforcement of copyright in the AI era. The Question of Fair Compensation Another major issue is whether creators should be paid when their works are used to train AI.Possible future directions include: • Licensing regimes for training data • Collective rights management models • New forms of remuneration tied to AI outputs This reflects a broader policy concern: ensuring that AI innovation does not undermine the economic foundation of creative industries. Broader Implications The developments in the EU are likely to have global impact:• The EU often acts as a regulatory standard-setter (the “Brussels Effect”) • Companies may adopt EU-compliant practices globally to avoid fragmentation • Other jurisdictions (e.g. UK, US) are watching closely At the same time, courts—particularly in the United States—are beginning to address similar issues, especially around fair use and AI training, suggesting a parallel global evolution in the law. Conclusion: A Turning Point The rise of generative AI has exposed deep flaws in existing copyright systems. The EU’s response signals a turning point:• From passive adaptation of old rules • To active redesign of copyright for the AI age Key themes—transparency, territorial reach, and fair compensation—are likely to define the next phase of legal development. For businesses, creators, and investors, this is not just a legal debate—it is a structural shift that will shape how AI is built, deployed, and monetised in Europe and beyond.
The report is significant because it goes beyond commentary—it proposes a reorientation of copyright law, with concrete expectations for: 1. AI Developers• Greater transparency about training data • Compliance with copyright rules across jurisdictions • Potential obligations to license content 2. Rights Holders• Strengthened ability to control or reserve rights • Improved tools to detect and enforce infringement 3. Legislators• A need to update copyright frameworks to address AI specifically • Clarification of how existing exceptions apply (or don’t apply) to GenAI A Shift in Territoriality: “Follow the Market” One of the most important emerging ideas is a shift in how copyright jurisdiction works. Traditionally, copyright law is territorial—linked to where an act (like copying) occurs. But AI complicates this:• Training may occur outside the EU • Outputs are used within the EU market The proposed shift suggests that EU copyright rules may follow the market rather than the location of training. This would mean:• Non-EU AI providers could still be subject to EU copyright law • Non-compliant systems could be excluded from the EU market This represents a major expansion of the EU’s regulatory reach. Transparency as a Central Requirement A key theme is transparency. Currently, creators often cannot determine:• Whether their work was used in training • How it was used • Whether outputs compete with their work The Parliament is pushing for:• Disclosure of training data sources (at least in summary form) • Mechanisms for rights holders to identify and assert claims Transparency is seen as a prerequisite for any meaningful enforcement of copyright in the AI era. The Question of Fair Compensation Another major issue is whether creators should be paid when their works are used to train AI.Possible future directions include: • Licensing regimes for training data • Collective rights management models • New forms of remuneration tied to AI outputs This reflects a broader policy concern: ensuring that AI innovation does not undermine the economic foundation of creative industries. Broader Implications The developments in the EU are likely to have global impact:• The EU often acts as a regulatory standard-setter (the “Brussels Effect”) • Companies may adopt EU-compliant practices globally to avoid fragmentation • Other jurisdictions (e.g. UK, US) are watching closely At the same time, courts—particularly in the United States—are beginning to address similar issues, especially around fair use and AI training, suggesting a parallel global evolution in the law. Conclusion: A Turning Point The rise of generative AI has exposed deep flaws in existing copyright systems. The EU’s response signals a turning point:• From passive adaptation of old rules • To active redesign of copyright for the AI age Key themes—transparency, territorial reach, and fair compensation—are likely to define the next phase of legal development. For businesses, creators, and investors, this is not just a legal debate—it is a structural shift that will shape how AI is built, deployed, and monetised in Europe and beyond.