The US Copyright Office has published new guidelines on copyright protection for works created with the help of generative AI. The document states that, although generative AI is a new technology, existing copyright principles can be applied without changes to the law.
The document notes that regardless of the complexity of the prompt or the number of its modifications, the final result reflects the user’s acceptance of the AI system’s interpretation, not authorship of the expression it contains. For example, the painting “Théâtre D’opéra Spatial,” created with Midjourney, cannot be registered as copyright-protected due to the lack of human authorship.
The Office also clarifies that using AI as a tool to support human creativity does not necessarily deprive such works of copyright protection. If AI is used to create a book outline or ideas for songs, this should not affect the ability to obtain copyright for the final product created by a human. However, AI-generated elements themselves are not subject to protection.
Protection can also be granted for works that include AI-generated content, provided there are significant creative modifications. For example, a comic book with images created by AI can be protected if a human organizes these images and adds text. Likewise, a film that includes special effects or background images created by AI can be protected, even if these elements individually are not eligible for protection.
The Copyright Office plans to continue researching the legal aspects of AI use, including issues related to training AI models on copyright-protected works. The next report on this issue is expected in July 2024.
