The U.S. Department of Justice filed a 20 page statement of interest in the Southern District of New York arguing that training large language models on copyrighted text constitutes fair use. The brief, supporting OpenAI in its legal battle with The New York Times, describes the training process as "extraordinarily transformative" and vital to U.S. national security. The filing warns that broad copyright liability or blanket licensing requirements would create barriers for smaller AI developers and give a competitive advantage to foreign adversaries.
Commerce Secretary Howard Lutnick separately urged G20 nations to embrace similar fair use standards for AI training to protect innovation while supporting creators. This intervention marks Washington's first formal move in the series of copyright lawsuits facing the AI industry. The DOJ argument focuses specifically on the copying stage of training, separating the act of learning statistical relationships from the acquisition of data and the generation of specific outputs.
Key sources
- SOURCE@wallstengine“Commerce Secretary Howard Lutnick urged G20 countries to embrace “fair use,” allowing AI companies to train on creators’ work”x.com
- SUPPORT@wallstengine“warning that restricting AI training based on a narrow reading of fair use could hurt U.S. AI leadership and innovation”x.com
- SUPPORT@andrewcurran_“the United States has a strong interest in 'rejecting any argument that training LLMs on copyrighted texts violates copyright law'”x.com
- SUPPORT@andrewcurran_“the creative possibilities and public benefits that LLM training advances far outweigh any competitive harm”x.com
- SUPPORT@mikeisaac“Trump admin argues it is in the 'national interest' to support OAI in the copyright case”x.com
- SUPPORT@rohanpaul_ai“This is Washington’s first formal intervention in the wider wave of copyright lawsuits over AI training”x.com
- SOURCE@adameisgrau“mandating training data licensing would be "legally incorrect"”x.com
- SUPPORT@anupamchander“rejects Copyright Office's pre-publication theory of "market dilution" for AI outputs”x.com