U.S. Copyright Office Explores Copyright Implications of Generative AI Training

Summary:

Le Bureau américain des droits d’auteur a publié une version préliminaire de son rapport sur l’intelligence artificielle (IA) concernant l’entraînement des modèles génératifs. Ce rapport explore les implications de l’utilisation des œuvres protégées par le droit d’auteur pour former des modèles d’IA et examine la défense de l’usage loyal. Bien que certaines utilisations puissent être considérées comme transformatrices, le rapport souligne que des facteurs, tels que le préjudice commercial, peuvent compliquer l’application de cette défense, avertissant des risques significatifs sur le marché des œuvres protégées.

Original Link:

Link

Original Article:

The U.S. Copyright Office released a prepublication version of Part Three of its artificial intelligence (AI) report addressing generative AI training. The report considers both whether training a generative AI model (and the models themselves) may involve acts that constitute prima facie copyright infringement, as well as potential fair use defenses. While recognizing that training models may often be transformative, the report notes that various factors may narrow a finding of transformativeness and cites a number of other potential obstacles to a finding of fair use, including market harm. The Copyright Office does not recommend any kind of statutory licensing or other interventions at this point.
In May 2025, the U.S. Copyright Office released a prepublication version of Part Three of its artificial intelligence (AI) report addressing generative AI training. After examining how copyrighted works are used in the development of generative AI systems, the report discusses whether those uses may present prima facie cases of copyright infringement, the applicability of the fair use defense to AI training, and potential licensing models for training data for use with generative AI, if required. The report’s conclusions indicate that creating a training dataset implicates reproduction rights, and the use of copyrighted works in retrieval-augmented generation (RAG) may also constitute prima facie infringement.
Defenses such as the fair use defense come into play as well, which protects certain beneficial uses of a work for criticism, commentary, news reporting, teaching, scholarship, or research, analyzed through four factors. The report notes that the commercial nature of the use and the potential for market harm are critical in determining the applicability of fair use. Ultimately, the Copyright Office concludes that copying for AI training poses significant risks to the market for copyrighted works, and barriers to licensing may impact fair use determinations. The long-term implications of the report are still to be determined due to pending developments in case law and potential legislative action.

Click to rate this post!
[Total: 0 Average: 0]

Comments

No comments yet. Why don’t you start the discussion?

Leave a Reply