TLDR: European Union lawmakers are preparing to advocate for the extraterritorial application of EU copyright laws to generative AI systems. This means AI models trained anywhere in the world would need to comply with EU copyright regulations if they are made available in the Union market. The proposed measures also include mandatory transparency on training data and a shift from an ‘opt-out’ to an ‘opt-in’ system for rights holders.
In a significant move set to reshape the landscape for artificial intelligence developers, European Union lawmakers are poised to call for the extraterritorial application of the bloc’s copyright laws to generative AI systems. This initiative, highlighted in a draft compromise obtained by MLex on October 20, 2025, signals a strong intent to ensure that AI models, regardless of their training origin, adhere to EU copyright standards if they are to operate within the Union’s market.
The core of the proposed resolution asserts that ‘The principle of territoriality is to be construed in such a way that, when generative AI models and systems are placed or made available on the Union market, EU copyright law […] applies regardless of the jurisdiction in which the copyright-relevant acts underpinning the training of those genAI models and systems take place.’ This ambitious stance aims to close potential loopholes where AI models trained in jurisdictions with less stringent copyright protections could bypass EU regulations when deployed within the bloc.
Beyond extraterritoriality, the draft also seeks to introduce mandatory transparency requirements regarding the data used to train generative AI models. This would compel AI developers to disclose the copyrighted materials ingested by their systems. Furthermore, a pivotal shift is envisioned from the current ‘opt-out’ system, where rights holders must actively prevent their content from being used for AI training, to an ‘opt-in’ mechanism. Under an opt-in regime, AI developers would be required to obtain explicit permission from rights holders before utilizing their protected works in training datasets.
This push by EU lawmakers comes amidst growing concerns about the impact of generative AI on intellectual property rights and the future of creativity. Existing EU copyright law, particularly the Text and Data Mining (TDM) exceptions introduced by the DSM Directive, has been deemed insufficient to address the complexities posed by mass reproductions of copyrighted works for AI training. Critics argue that current opt-out mechanisms, such as metadata or robots.txt files, are technically inadequate and that fragmented national transpositions of Article 4 of the DSM Directive create significant legal uncertainty.
Also Read:
- AI-Generated “Slop” Content Expected to Proliferate on Google
- Africa Emerges as a Leader in Ethical AI Innovation
While the resolution itself is non-binding, it carries substantial political weight, signaling the European Parliament’s firm position on these contentious issues. It also aligns with the European Commission’s ongoing review of the EU Copyright Directive, which is heavily focused on the implications of AI technologies and could lead to targeted amendments. The proposed framework aims to balance innovation with the protection and fair remuneration of authors, potentially advocating for a European register of permissions and the use of traceability technologies to ensure compliance. The ultimate goal is to establish clear rules on input/output distinctions, harmonized opt-out mechanisms, transparency obligations, and equitable licensing models to safeguard Europe’s rich and diverse cultural heritage in the age of machine-generated creativity.
