The lawsuit The New York Times brought against OpenAI is one of the cases that could decide whether training AI on copyrighted work is legal, so how each side argues matters well beyond the two parties. OpenAI has now published excerpts from 20 friend-of-the-court briefs filed in its support, from law professors, research libraries, the Electronic Frontier Foundation, disability groups, startups and the United States government. For anyone following AI copyright, the value is in seeing which arguments the defence is building its case around.
The briefs press three claims. The first is that training is transformative: it uses copyrighted works for a fundamentally different purpose than the works were made for, and does not give the public substitute access to them. The second, and the one most worth watching, targets what the filings call market dilution. The Times' theory is that AI outputs compete with and devalue its journalism even without copying its words, and the briefs argue this is not the kind of harm copyright law has ever recognised. New works have always been free to reduce the sales or value of older ones without infringing, they say, so aggregate competition from non-infringing outputs should not count as market harm. The third claim is practical: narrowing fair use would hurt medical research, education, accessibility tools for people with disabilities and open-source AI, not just large model builders.
The disability and library brief makes that breadth concrete, arguing the case touches patients, educators, farmers and independent creators who depend on systems others build, including open-source ones. It is a deliberate framing, casting a restriction on training as a loss for the public rather than a win for authors.
What readers should hold onto is that this is advocacy, not news. OpenAI chose which excerpts to publish and, by design, left out every brief filed in support of the Times. A collection of arguments for one side tells you how that side wants the fight framed, not how strong its case is or how the court will rule.
Still, the selection is informative precisely because it shows the ground the defence has chosen. The fight is increasingly being fought over market dilution, the question of whether competition from AI outputs is a harm copyright should prevent. That is the argument anyone following these cases should watch, because if the courts accept or reject the dilution theory, it will shape far more than this one dispute. The Times, of course, has its own briefs making the opposite case, and none of that is visible here.