U.S. Justice Department backs a fair-use path for text training in New York Times v. OpenAI
The Justice Department filed a Statement of Interest September 1 arguing that copying written works for language-model training can be highly transformative and need not harm the original market when outputs do not substitute for the works. It also disputes treating a hypothetical AI-training license market as automatically decisive. The filing is advice to the court, not a ruling; it focuses on written works, separates acquisition, training and output questions, and does not decide music cases against Anthropic, Suno or Udio.
Read this story →