US Backs OpenAI: 20-Page Brief Defends LLM Training on Copyrighted Data

For AI builders, the DOJ's brief in NYT v. OpenAI removes a potential legal sword hanging over training pipelines: if fair use covers unlicensed training data, model development can continue without per-article licensing costs. The government's 'AI dominance is critical' language makes clear that shutting down LLM training is now framed as a national security risk.

On September 1, 2026, the US Department of Justice filed a 20-page amicus brief in the Southern District of New York supporting OpenAI in The New York Times Company v. OpenAI and Microsoft, arguing that training large language models on copyrighted material is "exceedingly transformative" and protected by fair use. The filing is the clearest signal yet that the Trump administration views copyright enforcement against AI developers as a threat to national competitiveness, not merely a private commercial dispute. The DOJ explicitly linked its position to an executive order President Donald Trump signed in 2025 and warned that "constraining LLM development under a misunderstanding of fair use doctrine would thwart such creative and scientific progress while hindering American prosperity and economic mobility."

“

On September 1, 2026, the US Department of Justice filed a 20-page amicus brief in the Southern District of New York supporting OpenAI in The New York Times Company v.

The underlying case began in late 2023, when The New York Times sued OpenAI and Microsoft alleging the companies used millions of Times articles without permission to train ChatGPT and other models. The Times has framed the case as a test of whether the "fair use" exception in US copyright law permits AI companies to copy and ingest copyrighted works into training datasets. The DOJ's brief attacks the Times's interpretation directly, stating that "the creative possibilities and public benefits that (large language model) training advances far outweigh any competitive harm," and characterizing the Times's theory as creating an anti-competitive entry barrier that would "function primarily as large subsidies for old mainstream media companies."

The DOJ's intervention follows a pattern of court decisions favorable to AI developers. Last year, Judge William Alsup ordered Anthropic to pay a $1.5 billion copyright settlement to a group of writers whose works were used to train the company's AI, but that settlement was itself a negotiated resolution rather than a sweeping judicial finding of infringement. The government's brief could influence judicial thinking by lending the weight of the executive branch to the "transformative use" argument, though it is not binding on the court. It also signals to other publishers, authors, and creators that the federal government will oppose expansive copyright liability for AI training data, potentially discouraging further litigation or pushing plaintiffs toward licensing settlements.

For market participants, the brief reduces a key legal overhang for OpenAI and its primary investor Microsoft, which is named as a defendant. A finding for the Times could have set a precedent requiring AI developers to negotiate licenses with every rights holder whose work appears in training data, an outcome that would significantly raise costs and slow model iteration. By contrast, a ruling endorsing the DOJ's position would likely accelerate investment in AI infrastructure and model development, while putting pressure on traditional media companies to adapt their revenue models. The New York Times Company (NYSE: NYT) and Microsoft (NASDAQ: MSFT) both have exposure, though Microsoft's AI exposure is far larger than any potential liability from this case.