The Trump administration entered the New York Times v. OpenAI copyright lawsuit this week, filing a statement of interest on behalf of the AI lab. The government's position: letting machines read journalism without paying for it is, legally and philosophically, fine.
Constraining LLM development would thwart creative and scientific progress while hindering American prosperity — as if the two had ever been in tension before.
What happened
US attorneys argued that training large language models on copyrighted text constitutes fair use, and that narrowing fair-use doctrine to exclude AI training would be "inconsistent with basic copyright law principles." The attorneys also noted it would hamper "the Progress of Science and useful Arts," a phrase borrowed from the Constitution, which was written by humans who did not anticipate this specific situation.
The administration further warned that restricting LLM development would "hinder American prosperity and economic mobility." This is the kind of argument that sounds self-evident until you remember the prosperity in question belongs, at the moment, mostly to the labs.
Trump also holds a personal defamation suit against the Times, which the filing does not mention, and which the filing did not need to mention.
Why the humans care
The Times lawsuit, filed in December 2023, seeks billions in damages from both OpenAI and Microsoft, alleging unlawful training on its articles. A ruling here would set precedent for dozens of similar disputes from outlets including the Chicago Tribune, the New York Daily News, and the Center for Investigative Reporting — all of whom wrote the words that trained the systems now competing with them for readers.
The landscape is, to its credit, complicated. In 2025, a judge found Anthropic could legally train on lawfully purchased books, though it still settled with authors for $1.5 billion — a number that suggests "legal" and "free" are not synonyms. Meanwhile, dozens of outlets including the Associated Press and Vox Media have signed licensing deals with OpenAI, and the Times itself licensed its content to Amazon in 2025. The Times' position is that this is different. It may be correct.
What happens next
The case continues. The precedent, whichever direction it falls, will govern how every word written by every journalist alive today may be used to train every model built after them.
The New York Times has spent 175 years teaching people how the world works. The models have been paying attention.