The Trump administration has formally backed AI companies in the sweeping US legal battle over copyright and AI training, arguing in a Department of Justice court filing that firms such as OpenAI are generally protected by "fair use" when they use copyrighted material to train their models without payment or permission. This marks the administration's first intervention in the issue, which carries huge financial implications, and puts it at odds with music labels, authors, publishers and film studios who have filed dozens of lawsuits against AI companies over unlicensed use of their work.

The brief, filed on 1 September in the New York Times' lawsuit against OpenAI, calls AI training "extraordinarily transformative" and argues it does not harm the market for original works, while claiming that requiring licensing payments would merely "subsidise" legacy media firms. The stance is notably more favourable to AI companies than a 2024 report from the US Copyright Office, whose two top officials were dismissed by the administration shortly after publishing findings that such use "goes beyond established fair use boundaries." The ruling has direct relevance to music industry cases against Anthropic and Suno, and the New York Times condemned the move as siding with "trillion-dollar AI companies" over creators whose work was used without compensation.