United States of America: United States Copyright Office released pre-publication version of Part 3 of its Report on Copyright and Artificial Intelligence relating to Generative AI Training

On 9 May 2025, the United States Copyright Office released a pre-publication version of Part 3 of its report on Copyright and Artificial Intelligence, focusing on Generative AI Training. The report examines the use of copyrighted works in the development of generative AI systems, offering both technical and legal analysis. It discusses machine learning processes, generative language models, the characteristics and acquisition of training data, phases of model training, memorisation risks, where models may reproduce protected content and system deployment. Legally, it analyses potential prima facie infringement and the applicability of the fair use doctrine, addressing each of the four statutory factors in detail. The report also evaluates the feasibility and policy implications of licensing frameworks, offering an approach to assess relevant legal and policy issues and inform future regulatory considerations.