Amid Approval of EU AI Act, Creators Demand Stronger Protections for Rightsholders

“The concerns of the EU’s creator communities appear reasonable given the AI Act’s expectations that AI platform operators will themselves make their training dataset documentation compliant with EU copyright regulations.”

EU AI Act On March 13, the European Parliament approved the Artificial Intelligence (AI) Act, a major piece of legislation that lays the legal foundation of the European Union’s (EU) regulation of AI platforms. While the 459-page bill addresses some of the copyright and other intellectual property (IP) issues related to generative AI, European creator groups have called upon the EU’s parliamentary body to create more meaningful mechanisms for IP rightsholders to prevent their works from being incorporated into AI platform training models. Further, questions have been raised regarding the extraterritorial impact of reporting requirements and how they might implicate the development of copyright law in foreign jurisdictions.

Most of the provisions of the AI Act are aimed at protecting EU citizens from the worst safety and security risks that have become associated with the use of AI systems. The AI Act bans several uses of AI technologies, including the untargeted scraping of images to create facial recognition databases, emotion recognition in workplace and school settings, and social scoring systems. For EU law enforcement, the AI Act forbids predictive policing and also bans real-time biometric identification (RBI) applications with narrow exceptions for missing persons investigations and terrorist attack prevention. The newly approved AI regulatory framework also creates transparency and oversight obligations on high-risk AI applications in sectors like critical infrastructure, banking and election systems.

EU Creative and Cultural Organizations Want to Help Draft IP Enforcement Frameworks

Although the EU’s AI Act acknowledges existing legal frameworks protecting IP rightsholders, the legislative text does very little to establish new rules for the IP issues created by the AI Act’s reporting requirements. For example, general-purpose AI models are required by the AI Act to report data used in their training models, which must be documented “[w]ithout prejudice to the need to respect and protect intellectual property rights and confidential business information or trade secrets in accordance with Union and national law.” High-risk AI applications are similarly required to draft oversight documentation “without compromising their own intellectual property or trade secrets.”

The short shrift paid by the AI Act to IP rights enforcement has led to calls by the EU’s community of creators for the European Parliament to go much further in addressing the concerns of IP rightsholders. The same day that the AI Act was approved by the European Parliament, a broad coalition of 17 organizations representing interests in the EU’s creative and cultural sectors released a joint statement welcoming the AI Act’s approval but seeking further action.

While the obligations upon general-purpose and high-risk AI applications “provide a first step for rightsholders to enforce their rights,” these creators’ organizations asked the European Parliament to include their perspective in the drafting of IP enforcement frameworks surrounding the AI Act’s documentation and reporting requirements. Among the signatories to the joint statement include the International Confederation of Societies of Authors and Composers (CISAC), the International Federation of the Phonographic Industry (IFPI), and the International Federation of Film Producers Associations (FIAPF).

The concerns of the EU’s creator communities appear reasonable given the AI Act’s expectations that AI platform operators will themselves make their training dataset documentation compliant with EU copyright regulations. The AI Act acknowledges that techniques for training AI models in general-purpose applications may involve the retrieval and analysis of content that may be copyright-protected, and that such use requires the authorization of the relevant IP rightsholder.

Tensions Between EU’s Copyright Rules and U.S. Fair Use Doctrine Could Arise

Further questions have been raised about the extraterritorial reach of the AI Act’s reporting requirements on AI businesses that are headquartered in jurisdictions with different copyright regimes. Clause 106 of the AI Act provides that general-purpose AI companies must ensure that their reporting obligations are completed in compliance with Article 4 of the EU’s 2019 Copyright Directive, which creates an exception to EU copyright law for text and data mining used for scientific research. Companies must comply with these provisions of the EU’s Copyright Directive “regardless of the jurisdiction in which the copyright-relevant acts underpinning the training of those general-purpose AI models take place.”

The AI Act states that such a broad application of the Copyright Directive on acts occurring in non-EU companies is necessary to ensure a level playing field so that no AI company operating in the EU can benefit by applying lower copyright standards from their home country. Commentators have pointed out that such a legal framework creates friction with at least United States’ case law surrounding fair uses of copyrighted content for training AI models. The spectrum of viewpoints on applying U.S. fair use doctrine to AI model training was on display at a Senate IP Subcommittee hearing last July, with executives from Adobe and Simplicity AI discussing very different interpretations of the fair use defense in that context.

As of late February, there are several major copyright infringement lawsuits winding their way through various U.S. district courts, especially in the Southern District of New York and Northern District of California. Lower court rulings on fair use defenses expected to be raised in those cases are likely to be addressed by circuit courts that will issue appellate rulings in the coming years, developing fair use doctrine in the generative AI context over time. It’s possible that the new reporting obligations on AI companies without regard to jurisdiction could spawn further infringement litigation in the U.S. and elsewhere. This January, a commission of the French National Assembly recommended amending the EU’s Copyright Directive to establish an international AI treaty regarding copyright issues in generative AI, which would go much further to establish the EU’s leadership on regulating AI and copyright matters during these early years of generative AI.

Image Source: Deposit Photos
Author: phonlamai
Image ID: 238275390 

Share

Warning & Disclaimer: The pages, articles and comments on IPWatchdog.com do not constitute legal advice, nor do they create any attorney-client relationship. The articles published express the personal opinion and views of the author as of the time of publication and should not be attributed to the author’s employer, clients or the sponsors of IPWatchdog.com.

Join the Discussion

One comment so far.

  • [Avatar for Anon]
    Anon
    March 17, 2024 06:52 pm

    The ultra vires extra-territorial portions will fail.

    As they must.