AI and Copyright Rights - Third Quarter Report
In the dynamic world of technology, the intersection of copyright law and generative AI (GAI) is a topic of ongoing debate. This article will explore recent developments in this area, focusing on significant court cases and legislative initiatives.
EU and China Lead the Way in AI Regulation
The European Union has taken a proactive approach, publishing the General Purpose AI Code of Practice, which includes guidelines on copyright. Signatories are required to implement a copyright policy and keep it up-to-date. On the other hand, China has ruled in favour of the defence in a GAI copyright infringement case, finding that a work generated using AI did not involve enough human creativity to be protected.
US Courts Weigh in on Fair Use
In the United States, the use of copyright-protected works to train GAI models remains legally unsettled but is evolving. Recent court decisions, such as Bartz v. Anthropic and Kadrey v. Meta, have suggested that using copyrighted materials for AI training constitutes a "transformative" use, which is a key factor supporting a fair use defense under U.S. copyright law. However, fair use is fact-specific, and courts will consider the exact nature and consequences of training on copyrighted works on a case-by-case basis.
The U.S. Copyright Office, in its report on GAI and copyright, stated that copyright may exist when AI is used as a tool to support human creativity but not when works are essentially autogenerated. They also concluded that there is no immediate need for Congress to create a compulsory license scheme; voluntary collective licensing could suffice to meet demand.
Global Diversity in AI Copyright Regulation
Globally, the legal environment is similarly unsettled and varies by jurisdiction. Denmark has proposed new deepfake legislation that aims to give individuals rights in their own bodies, facial features, and voices, marking the first of its kind in Europe. In contrast, the UK's Data (Use and Access) Bill, which does not include proposed amendments that would have substantially reformed copyright law for the AI age, has been approved by lawmakers.
Ongoing Legal Battles and Future Guidance
As of mid-2025, significant legal battles are ongoing, with Toronto Star Newspapers Limited et al. suing OpenAI for using their news content to train its GAI model, and Disney Enterprises Inc. suing Midjourney Inc. for allegedly infringing outputs from a GAI model trained with copyright-protected materials. The situation remains fluid, with guidance expected as more cases are adjudicated.
In conclusion, while the legal landscape for the use of copyright-protected works to train GAI models is evolving, it remains a complex and diverse field. As more cases are decided and guidance is issued, we can expect a clearer understanding of the boundaries between AI and copyright.
[1] Bartz v. Anthropic PBC [2] U.S. Copyright Office Report on Generative AI and Copyright [3] U.S. Copyright Office's Part 2 Report on Generative AI and Copyright [4] Global Debate on AI and Copyright [5] Kadrey v. Meta Platforms Inc.
- The European Union's General Purpose AI Code of Practice requires signatories to implement a copyright policy and keep it updated, marking a proactive approach in the intersection of copyright law and AI.
- In the United States, court decisions like Bartz v. Anthropic and Kadrey v. Meta suggest that using copyrighted materials for AI training may fall under the 'transformative' category that supports a fair use defense, but the legal landscape remains uncertain due to its fact-specific nature.
- Globally, the legal environment for AI and copyright regulation varies greatly, with countries such as Denmark proposing deepfake legislation and the UK's Data (Use and Access) Bill passing without substantial reform, leading to a complex and diverse field in this area.
