New EU Code of Practice for General-Purpose AI Models: What impact will it have on copyright?

New EU Code of Practice for General-Purpose AI Models: Implications for Copyright

The European Union (EU) has recently rolled out a new Code of Practice designed to regulate general-purpose artificial intelligence (AI) models. This initiative is part of the EU’s larger effort to promote responsible development and deployment of AI technologies, with a strong focus on ethical considerations and legal frameworks. A significant aspect of this code is its approach to copyright, which raises important questions for creators, developers, and users of AI-generated content.

Background

The emergence of general-purpose AI models, like OpenAI’s GPT series and Google’s Bard, has revolutionized various sectors by enabling the creation of text, images, and other media. However, the swift advancement of these technologies has left existing legal frameworks, especially in copyright, struggling to keep pace.

In April 2023, the European Commission proposed a comprehensive regulatory framework for AI, which included this new Code of Practice. The goal is to enhance transparency, accountability, and safety in AI systems while tackling potential copyright infringement issues.

Key Features of the Code

The Code of Practice introduces several important provisions that directly relate to copyright:

1. Transparency Requirements: AI developers are now required to disclose the datasets used to train their models. This transparency is vital for assessing the originality of generated content and addressing copyright claims.
2. User Responsibilities: Those who use AI-generated content are encouraged to ensure their usage aligns with copyright laws. This involves verifying the originality of the content and securing necessary licenses when needed.
3. Fair Use and Licensing: The code highlights the significance of fair use in AI-generated works, allowing for certain exceptions in copyright law that could benefit both creators and users.
4. Monitoring and Reporting: Developers must establish systems to monitor the use of their AI technologies to prevent copyright infringement, including mechanisms for users to report any violations of their rights.

Implementation Timeline

The Code of Practice was officially adopted in September 2023, with a phased approach to implementation:
– Q4 2023: Initial guidelines and best practices will be released for developers and users.
– 2024: All AI developers operating within the EU will be expected to fully comply with the code.
– 2025: The European Commission plans to review the code’s impact and make necessary adjustments based on stakeholder feedback.

Copyright Implications

The introduction of this Code of Practice is poised to have several important implications for copyright:

• Increased Clarity: By offering clear guidelines on dataset usage and AI-generated content, the code aims to clarify issues surrounding copyright ownership and infringement.
• Empowerment of Creators: Creators could see enhanced protections against unauthorized use of their work, as the code promotes accountability among AI developers and users.
• Legal Precedents: As the code is put into practice, it may establish new legal precedents regarding the copyright status of AI-generated works, particularly concerning whether such works can be copyrighted at all.
• Innovation and Collaboration: The focus on fair use may encourage innovation by allowing creators to build on existing works without fear of legal repercussions, as long as they adhere to fair use guidelines.

Conclusion

The new EU Code of Practice for General-Purpose AI Models represents a significant advancement in addressing the intricate relationship between AI technologies and copyright law. As the code is implemented, its effects on creators, developers, and users will become increasingly evident, shaping the future of copyright in the AI era. Stakeholders will need to navigate these changes thoughtfully to ensure compliance while promoting creativity and innovation in the digital landscape.