The Justice Department has backed OpenAI and Microsoft in their high-stakes copyright fight with The New York Times, arguing that training AI models on copyrighted works constitutes fair use and that restricting the practice could give foreign rivals an advantage.
In a 20-page statement filed Sept. 1 in Manhattan federal court, the Trump administration weighed in on one of the most closely watched AI copyright disputes in the country. The DOJ urged Judge Sidney H. Stein to reject the Times’ argument that training large language models on copyrighted texts violates copyright law.
“The United States has a strong interest in this court rejecting any argument that training LLMs on copyrighted texts violates copyright law,” the filing states, citing concerns ranging from scientific advancement to national security.
Associate Attorney General Stanley Woodward wrote on X: “AI dominance is critical to promote national security, prosperity, and economic mobility for all Americans. This Administration will never let our Nation be at a disadvantage relative to our foreign adversaries based on a plainly incorrect understanding of copyright law.”
‘Extraordinarily transformative’
The government argued that AI training is highly transformative, an important consideration in the fair-use analysis. While the Times alleges OpenAI and Microsoft used millions of its articles without permission to build ChatGPT, the DOJ contends that copying works to teach an LLM statistical patterns differs fundamentally from reproducing the works themselves.
The filing drew an unusual parallel: when a teenage Joan Didion typed out Ernest Hemingway’s stories “to learn how the sentences worked,” she wasn’t infringing copyright. The government used the example to argue that learning from copyrighted works should be distinguished from reproducing those works.
“Constraining LLM development under a misunderstanding of fair use doctrine would thwart such creative and scientific progress while hindering American prosperity and economic mobility,” the brief reads.
The publisher’s response
In a statement to The Washington Post, The Times spokesperson, Graham James, accused the administration of “siding with a handful of trillion-dollar AI companies at the expense of the countless American creators whose work they stole.”
“Both AI and creators can thrive — AI companies simply need to pay fairly for the content that makes their products possible, as copyright law requires,” James said.
Commerce Secretary Howard Lutnick separately told G20 officials that nations should embrace fair use while finding ways to “protect artists.”
More must-read AI coverage
- SS&C Intralinks DealCentre AI vs. Datasite: Which platform is built for the future of dealmaking?
- SS&C Intralinks FundCentre AI vs. Juniper Square: Which platform better supports modern private markets fund managers?
- Why Data, Not Models, Determines AI Success
- The Rise of the AI-Native Factory: How Physical AI Is Transforming Manufacturing
The competition question
The filing’s most striking argument concerns market dynamics. The DOJ warned that requiring licensing fees would create an AI oligopoly, benefiting only the wealthiest tech companies while “disproportionately benefit legacy media outlets” with the largest archives.
“It is not in the public’s interest for the largest technology companies to have an oligopoly on LLM training due to licensing entry barriers that function primarily as large subsidies for old mainstream media companies,” the government wrote.
The legal question is far from settled
The government’s intervention does not settle whether AI training is fair use. Courts have reached different conclusions in related cases.
In the Anthropic litigation, a judge found that using books to train its AI models was fair use but ruled that Anthropic could not rely on fair use to justify pirated copies it acquired and retained separately. Meta, meanwhile, prevailed in a separate copyright case, although the judge stressed that the ruling was based on the evidence presented and did not establish that AI training is always fair use.
That uncertainty means the New York Times case remains important. A ruling could influence whether AI developers must negotiate licenses for more of the material used to train their models, potentially affecting development costs, competition among AI companies and the licensing market for publishers. For businesses adopting generative AI, the outcome could ultimately shape both the cost and availability of future models.