Trump Administration Backs OpenAI in Copyright Lawsuit

The Trump administration has intervened in the high-stakes legal battle between OpenAI and the New York Times, filing a formal letter on Tuesday that urges the court to reject the publisher's copyright claims. Submitted to U.S. District Judge Sidney H. Stein, the Department of Justice's letter argues that restricting large language model development under a narrow view of fair use would "thwart such creative and scientific progress" and harm American economic competitiveness. The government asserted that training AI models is highly transformative and does not directly compete with the original articles.

The New York Times originally sued OpenAI and Microsoft in 2023, alleging unauthorized use of its articles. NYT spokesperson Graham James criticized the government's intervention, stating the administration is favoring tech giants at the expense of creators. Mary Rasenberger, CEO of the Authors Guild—which also sued OpenAI in 2023—expressed deep disappointment with the government's stance. However, intellectual property experts like Evan Brown of Neal & McDevitt and Pamela Samuelson of the Berkeley Center for Law & Technology noted that while Judge Stein is not bound by the letter, federal input carries immense weight and aligns with recent judicial trends.

This intervention comes amid a wave of industry-defining litigation. In the Kadrey v. Meta case, Meta secured a technical victory when a judge found insufficient evidence of training harm. Conversely, Anthropic was ordered to pay $1.5 billion in damages—the largest copyright settlement in U.S. history—after a judge ruled that while its training of the Claude model constituted fair use, the company had pirated the underlying books. Anthropic also faces newer lawsuits from Sony, Warner Music, and Universal Music Group over Claude's training data.

For AI practitioners and developers, the federal government's explicit support of the fair use defense provides a significant psychological and regulatory boost. If the courts adopt the administration's view that training LLMs is inherently transformative, it will drastically lower the legal risks and potential licensing costs associated with compiling massive datasets. This could prevent a chilling effect on AI startups, ensuring that developers can continue training advanced models without the constant threat of existential copyright litigation.