In the Trump administration OpenAI lawsuit fight, the Justice Department has filed a statement of interest in Manhattan federal court backing OpenAI’s fair-use defense against The New York Times. The filing gives the administration’s policy weight to a central argument in the case: that using copyrighted text to train a large language model can generally qualify as fair use. It does not settle the lawsuit.
Reuters reported that the government’s brief says the United States has an interest in rejecting the view that training large language models on copyrighted text violates copyright law. The Justice Department describes the training process as “extraordinarily” transformative and says excessive restrictions could hurt scientific development, national security, economic mobility and the United States’ position in AI.
According to Politico’s account of the filing, the department argues that model training and a chatbot’s output should be assessed separately. Its theory is that training uses works to identify statistical patterns in language, rather than to reproduce the expressive content of a particular article. That is OpenAI’s legal theory with the federal government now standing behind it, not a court finding.
Does the Justice Department filing decide the New York Times lawsuit?
No. A statement of interest is advisory, Reuters reported. Judge Sidney H. Stein is not required to adopt the government’s view or rule that OpenAI’s conduct was fair use. Reuters said the filing could nevertheless strengthen technology companies’ arguments in this and related cases. It also reported that the intervention appears to be the first time the U.S. government has weighed in during the current wave of lawsuits over AI training.
The Times sued OpenAI and Microsoft in 2023, alleging the companies used millions of the newspaper’s articles without permission to train OpenAI systems, including the technology behind ChatGPT. OpenAI has denied wrongdoing and has maintained that training on publicly available material is protected by fair use, according to TheWrap.
The Times rejected the administration’s position. Graham James, a spokesperson for the newspaper, said the government was siding with large AI companies over creators and argued that companies should pay for the material used to make their products. That is the publisher’s contention; the court has yet to resolve it.
What will the court consider in the fair-use dispute?
Fair use is a case-specific doctrine, not a blanket permit for unlicensed copying. The U.S. Copyright Office says courts weigh four factors: the purpose and character of the use, the nature of the copyrighted work, the amount and importance of what was used, and the effect on the work’s potential market or value.
That framework explains the government’s emphasis on transformation and competition with Times journalism. The administration is asking the court to accept its view of those issues in the context of LLM training. The Times argues that unlicensed use threatens the economic basis for human-created reporting. Reuters reported that the first two judicial decisions addressing AI-training fair use reached different results, so neither side has a clean, universal answer waiting on the shelf.
This story draws on original reporting from The Verge.
