The Justice Department backs OpenAI against The New York Times

The US government has told a judge that OpenAI can train on The New York Times, and that stopping it would hurt the country.

The filing

In court papers filed on September 1, 2026, the Justice Department backed OpenAI’s fair use defense in the Times’ copyright lawsuit. The filing says “the creative possibilities and public benefits” of training large language models “far outweigh any competitive harm.” It argues that a decision for the Times would hinder “American prosperity and economic mobility,” and that models trained this way help national security officials analyze real-world facts.

The Times sued OpenAI and Microsoft in 2023 over millions of articles it says were used without permission, some of which ChatGPT reproduced almost verbatim. Its reply this week was short: “AI companies simply need to pay fairly for the content that makes their products possible, as copyright law requires.”

Why the government’s weight matters

Fair use for AI training has not been settled at this level, and the executive branch putting its position on the record changes the odds for every publisher suing a model maker. Last week Sony and Warner sued Anthropic over song lyrics, the mirror case on the music side. The administration’s argument links training to national security and economic growth rather than to the economics of publishing, which is the frame the courts will now have to weigh.

What comes next

The filing does not decide the case; it tells the court which way the government thinks the balance should tip. For publishers, the practical effect is that the licensing deals some of them have already signed with AI companies look more like the ceiling than the floor. For OpenAI, the risk of a training-data injunction just got smaller, and the price of settling got lower.