> ## Content Index
> Fetch the complete content index at: https://spyrigend.ghost.io/llms.txt
> Use this file to discover other available public pages before exploring further.

# Justice Department tells court AI training on news articles is fair use
- URL: https://spyrigend.ghost.io/justice-department-tells-court-ai-training-on-news-articles-is-fair-use/
- Published: 2026-09-03T03:49:39.000Z
- Updated: 2026-09-03T03:49:39.000Z
- Description: DoJ backs OpenAI vs. The New York Times: training AI on copyrighted works is fair use, not infringement.
- Author: Quinn Hillerich
- Tags: All Articles, News, AI

DOJ backs OpenAI and Microsoft against the New York Times, arguing licensing mandates would entrench the largest tech firms and hand foreign AI rivals an edge.

The Justice Department filed a statement of interest late Tuesday in a Manhattan federal court arguing that training large language models on copyrighted news articles is protected fair use. The filing supports OpenAI and Microsoft in a consolidated case brought by the New York Times, Ziff Davis (CNET's parent company), and roughly 400 local newspapers, now before Judge Sidney Stein in the Southern District of New York.

## DOJ's core argument

The government's brief separates training from output. Copying text into a model, it argues, converts that text into statistical patterns rather than republishing it as expression, so training should be evaluated on its own terms rather than folded into claims about what a chatbot later generates. Associate Attorney General Stanley Woodward said AI dominance is critical to national security, prosperity, and economic mobility for all Americans, and that a ruling against fair use would represent an incorrect reading of copyright law.

The filing also warns that requiring licenses for training data would effectively subsidize legacy publishers and favor the tech companies with enough capital to pay for it, entrenching the largest players rather than opening the field. It argues that restricting access to training data domestically would cede ground to foreign AI developers not bound by the same rules.

## Case background

The Times sued OpenAI and Microsoft in December 2023, alleging ChatGPT and Copilot were trained on millions of its articles without permission or compensation. Ziff Davis filed its own suit in 2025, and authors including George R.R. Martin and John Grisham have brought related claims against AI developers over training data. Anthropic settled a separate authors' suit for $1.5 billion.

A New York Times spokesperson said the administration is siding with a handful of trillion-dollar AI companies at the expense of the American creators whose work they used, and said AI and journalism can both succeed if companies pay for the content they use.

Judge Stein has ordered the Times to explain in writing by September 11 why the case should not be stayed pending summary judgment rulings in other consolidated suits, with any response from the defendants due September 18.

## What it means

A statement of interest carries no binding legal weight, but courts weigh the government's position when deciding fair use questions. If Judge Stein or other courts adopt DOJ's framing, unlicensed training would become the default posture for AI companies operating in the U.S., though copying or reproducing a work in a model's output could still count as infringement. A ruling the other way would push AI developers toward negotiated licenses or force them to rebuild training sets without the disputed material, raising costs across the industry.

The case is one of several consolidated actions in the Southern District of New York, and how it resolves will likely shape licensing negotiations between AI companies and publishers well beyond this suit. Judge Stein's response to the government's filing, and whether he grants the September 11 stay request, will be an early signal of how much weight the court gives the administration's position.