AI Training and Copyright in Europe: A Potential Shift Beyond Territoriality
The European Parliament adopted a Resolution on Copyright and Generative Artificial Intelligence on March 10, 2026, signaling a potentially significant shift in European copyright policy. Although the Resolution itself is not legally binding, it reflects the growing view within EU institutions that traditional copyright rules may not adequately address the realities of generative AI training.
Key Developments
The Resolution suggests a broader territorial reach of EU copyright law for AI systems offered in the European market. A notable proposal includes a potentially retroactive flat rate licensing fee of 5 to 7% of global turnover to compensate the creative industry for the training of AI systems. If implemented, this approach could affect AI developers and deployers worldwide, including companies based in the United States, and may lead to tensions with non-EU legal regimes governing AI training practices.
AI Training and Existing Copyright Limits
European copyright law has historically developed in a technological environment very different from today’s digital economy. Its central objective has been to protect individual creative works against unauthorized acts of reproduction and distribution. However, generative AI systems operate fundamentally differently by processing extremely large datasets. Traditional copyright concepts may not adequately apply to AI training, as it relies on large-scale computational analysis rather than direct interaction with individual works.
The Economic Importance of Generative AI
Generative AI is becoming integral to modern business processes, producing text, software, visual content, and analytical insights. The legal sensitivity surrounding what content was used to train AI models raises vital questions about authorization and compensation. A central debate in EU discussions is whether the text-and-data-mining (TDM) exception in the 2019 EU Copyright in the Digital Single Market Directive covers generative AI training.
Emerging Policy Conclusions
The Committee Report suggests that the traditional copyright framework may not be sufficient for the AI era. This has led to a dual policy objective: regulating AI training while ensuring the economic sustainability of Europe’s creative industries. The Report proposes establishing an immediate, simple, flat-rate copyright fee to compensate for the use of data from European creatives.
Legal Uncertainty and Court Proceedings
The legal uncertainty surrounding AI training is evident in ongoing national court proceedings. Notably, a November 2025 decision by the Regional Court of Munich concluded that certain AI training forms could constitute copyright infringement. This contrasts with a UK ruling where the High Court dismissed a copyright claim against an AI model.
Proposals for Standardized Opt-out Mechanisms
The JURI Report discusses creating standardized opt-out mechanisms that allow rights holders to signal that their content should not be used for AI training. Such signals should be machine-readable, allowing AI developers to recognize them during data collection. Additionally, the Report suggests maintaining a centralized European register to track opt-out declarations.
Increased Transparency Requirements
The European Parliament emphasizes the need for AI developers to disclose more information about the copyrighted materials used in training. This transparency could involve communication to a public authority such as the EU Intellectual Property Office (EUIPO).
Challenges of Territoriality in Copyright Law
Traditionally, copyright law has been governed by the principle of territoriality, where protection is determined by local laws. The Resolution suggests that this principle may require reevaluation in the context of AI, proposing that EU copyright rules apply whenever an AI system is offered in the EU market, regardless of where the model was trained.
Implications for Global Technology Companies
Should the EU adopt this approach, it could create far-reaching implications for companies outside Europe. AI providers may find themselves subject to EU copyright requirements even when their operations are entirely outside the Union. This shift could lead to significant regulatory tensions with jurisdictions that maintain different views on AI training legality.
Practical Implications for Companies
While the European Parliament’s Resolution does not have immediate legal effect, it serves as an early indicator of future regulatory developments. Companies operating within the AI ecosystem should assess how their systems align with European copyright rules and ensure compliance with emerging transparency and opt-out requirements.
The evolving debate on territorial reach indicates that companies outside Europe may need to adapt their practices to avoid potential legal risks associated with AI services offered to European users.
Conclusion
The European Parliament’s discussions are gradually moving beyond traditional copyright law towards a broader concept of data governance for AI training, intersecting with copyright law, the EU AI Act, and data protection law. This evolving landscape will significantly impact how generative AI operates in a global context, necessitating careful consideration of compliance across differing regulatory frameworks.
For further information, refer to the Report of the European Parliament (JURI Committee), the Explanatory Statement of the JURI Committee, and the European Parliament Resolution.
