U.S. government backs OpenAI in New York Times copyright case

The Justice Department is urging a federal judge to side with OpenAI and Microsoft in The New York Times’ copyright lawsuit, arguing that training large language models on large collections of copyrighted writing can generally qualify as fair use.

The department filed a statement of interest Tuesday, September 1, 2026, in the Southern District of New York. The filing, reported publicly on September 2, puts the federal government behind a major legal argument in one of the most closely watched U.S. disputes over how artificial-intelligence systems are built.

The filing does not decide whether OpenAI or Microsoft violated copyright law. It is an advisory submission to the court, and the judge still must evaluate the parties’ evidence and legal claims.

What the Justice Department argued

Justice Department lawyers urged the court to reject a broad theory that using copyrighted news articles to train large language models is generally unlawful. The government characterized the process as highly transformative because models analyze large quantities of text to learn patterns and generate new material rather than simply republish each article as a conventional copy.

The department argued that the creative and public benefits of AI training outweigh the competitive harm alleged by publishers. It linked continued AI development to creative and scientific progress, economic mobility, national security and U.S. competitiveness.

The filing also warned that imposing broad copyright liability or licensing requirements could make it harder for smaller companies to compete in the large-language-model market. Those are the government’s arguments for the judge to consider, not findings that have already become law.

What The New York Times alleges

The Times has argued that OpenAI and Microsoft used millions of its articles without permission or payment to develop products including ChatGPT and Copilot. The newspaper says the systems can sometimes reproduce or closely summarize its journalism and may compete with the original reporting.

The Times is seeking monetary damages and court-ordered relief. The allegations remain contested. The court must weigh evidence about how the companies obtained and used the material, whether the use was sufficiently transformative, whether model outputs substitute for the original work and how the factors in copyright’s fair-use doctrine apply.

What the filing does — and does not — change

The Justice Department’s filing may influence the judge’s analysis and strengthen the negotiating position of AI companies in licensing disputes. It does not create a blanket exemption for training models on copyrighted works, and it does not eliminate copyright claims by publishers, authors, artists or other rights holders.

Even a ruling favorable to the government’s position would not necessarily resolve every AI copyright dispute. Cases may turn on the source material, the way it was collected and used, the model’s outputs, evidence of memorization and the effect on a creator’s market.

Why the case matters nationally

The outcome could affect whether AI developers need licenses or payments to use news, books, images, music and other protected material in training datasets. It could also influence the cost and availability of AI tools and the business models supporting journalism and other creative industries.

The case is part of a wider wave of litigation involving publishers, authors, artists, musicians and AI developers. Separate cases and settlements are testing questions about training data, copied outputs, licensing and the economic value of human-created work.

The U.S. Copyright Office is separately examining copyright questions raised by artificial intelligence, including the use of copyrighted works in AI training. The office says its report is being issued in multiple parts; its analysis provides policy context but is not binding law or a decision in the Times case.

What to watch next

The case is moving into a critical pretrial phase. The New York Times reported that Judge Sidney H. Stein gave the parties until Friday, September 11, 2026, to submit summary-judgment motions.

The major questions include whether training is sufficiently transformative, whether AI systems memorize or reproduce protected expression, whether their products substitute for original reporting and whether licensing should be required in particular circumstances.

Until the judge rules, the Justice Department’s position is best understood as an influential legal argument — not a final determination that AI training is lawful.

Sources

Look for updates to this story

Discover more from Interactive News

Subscribe to get the latest posts sent to your email.