OpenAI’s fair-use case would narrow publishers’ AI-training leverage
Amicus filings ask the court to separate competition from copyright harm in the market for AI training.
OpenAI published excerpts from amicus briefs in its copyright fight with The New York Times, arguing that model training is transformative and does not substitute for access to source works. One filing says competitive “market dilution” from noninfringing outputs should not count as copyright market harm, while another says newspaper sales were unaffected by training itself. If courts accept that framing, publishers would have less leverage to demand payment for training use and would need stronger evidence of substitution to sustain licensing claims.