Breaking

DOJ Backs OpenAI’s Fair-Use Position in New York Times Copyright Case

September 3, 2026 at 20:12 EDT

Policy & Regulation
Foundation Models

The U.S. Justice Department has urged a federal judge to reject arguments that training large language models on copyrighted text is inherently unlawful, siding with OpenAI in its copyright dispute with The New York Times. The government filed a 20-page statement of interest on September 1 in the consolidated litigation in the Southern District of New York, arguing that AI training serves transformative purposes and supports U.S. innovation and national security.

September 1, 2026 · U.S. Justice Department

AI training enters a defining copyright fight

The DOJ urged a federal court to reject the argument that training language models on copyrighted text is inherently unlawful, backing OpenAI in its dispute with The New York Times.

What the filing says

Training

Learning language patterns is different from reproducing the original work.

Competition

Broad licensing requirements could raise barriers for smaller AI developers.

Limit

Outputs that reconstruct protected works may raise separate legal questions.

A $1.5B precedent shows the distinction

Training: fair use

Book piracy: $1.5B

The legal question can change depending on how works are obtained, trained on, or reproduced.

Innovation case

Transformative learning

The government says models learn statistical patterns to create new tools and support scientific progress.

Creator case

Permission and payment

Publishers and authors argue that commercial AI systems should compensate the work that makes them possible.

Next question

Where does training end and infringement begin?

The court’s ruling could shape future disputes over training data, licensing, memorization, and generated outputs.

Continue reading

The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.