On Tuesday, the US government filed an amicus brief in a federal court in Manhattan - and it didn’t hedge. In the case between OpenAI and the New York Times, the Trump administration lined up behind the AI lab. Training large language models on copyrighted text generally does not violate copyright law, the brief says. It’s fair use. Reuters broke the story.
That matters because it’s the first time the government has stepped into this wave of lawsuits at all. Over the past two years, authors, publishers, music labels and news outlets have filed dozens of cases against OpenAI, Anthropic and Meta. The accusation is always the same: the models were trained on other people’s work, without asking and without paying.
What the brief says
The language is blunt. The US has “a strong interest in this court rejecting any argument that training LLMs on copyrighted texts violates copyright law” - grounded in scientific advancement and national security. AI training is “extraordinarily” transformative, the brief continues. And: language models are already helping researchers achieve breakthroughs across fields. A cramped reading of fair use would choke that progress and hurt American prosperity.
The case itself has been running since 2023. The Times accuses OpenAI and its major backer Microsoft of using millions of articles without permission to train ChatGPT. The brief was filed with Judge Sidney Stein in the Southern District of New York.
Why this touches Anthropic too
For us here on clauding.de, the interesting part isn’t OpenAI - it’s the aside. Reuters explicitly counts Anthropic among the companies facing dozens of these suits. Anthropic already settled a copyright case with book authors over the summer, and at the end of August music publishers filed a fresh one. If a court follows the government’s reasoning, it doesn’t just help OpenAI - it helps the whole industry.
An amicus brief is not a ruling. It carries advisory weight, not binding weight. The court doesn’t have to follow it. The first two judges to take up the fair-use question reached opposite conclusions last year, so the law here is anything but settled.
What gives me pause: here’s a government invoking national security to protect the training practices of private AI firms. That’s a strong signal of how important AI has become to the US. But the people whose work got trained on walk away with nothing, at least for now. The real question - whether and how creators get paid - the brief doesn’t answer. It only says: this isn’t illegal. For the people on the other side, that’s cold comfort.
Sources: GV Wire / Reuters: US Government Backs OpenAI in New York Times Copyright Case, Business Standard: US government backs OpenAI in New York Times’, others’ copyright case, gHacks: Trump Administration Backs OpenAI, Calling AI Training Fair Use