DOJ Backs OpenAI in Copyright Fight, Citing National Security

The future of the internet's next layer won't be decided by philosophers debating the nature of human creativity. It will be determined in a Manhattan federal courthouse by a judge applying a four-factor statutory test from 1976. That is the bottleneck where the artificial intelligence industry either secures its supply chain or faces an existential toll booth. And on September 1, 2026, the U.S. Department of Justice walked into that bottleneck, filing a 20-page Statement of Interest that placed the weight of the federal government on the side of the machines.[2][4]

The DOJ's intervention in The New York Times v. OpenAI marks the first time the administration has formally staked a position in the sprawling multidistrict litigation over AI training data. The government's argument is blunt: training large language models on copyrighted text is "extraordinarily transformative" and generally qualifies as fair use.[2][3]

But the Justice Department did not just cite legal precedent; it pulled the national security ripcord. According to the filing, imposing broad copyright liability on AI training would "severely hamper" technological progress. The brief argues that rules making it difficult to develop a robust AI industry "threaten national security and give a competitive advantage to foreign adversaries who are not so encumbered."[1][4]

The stakes are difficult to overstate. OpenAI, currently negotiating a funding round that could value the company at $852 billion, relies on scraping vast swaths of the open web to build models like GPT-4. If courts rule that ingesting that data requires a license, the financial math of generative AI changes overnight.[4]

The Times, which sued OpenAI and its primary backer Microsoft in December 2023, immediately pushed back. The newspaper alleges that millions of its articles were ingested without permission to build products that now compete directly with its journalism. Times spokesperson Graham James did not mince words, stating that the government is "siding with a handful of trillion-dollar A.I. companies at the expense of American creators."[1]

The DOJ's brief attempts to thread a delicate needle by separating the act of training a model from what the model actually spits out. Section 107 of the Copyright Act protects unlicensed use for purposes like research and scholarship. The government contends that copying text to teach a model how language works—developing generalized reasoning skills—is fundamentally different from reproducing the text for a reader.[1][3]

"The fair-use inquiry hinges on the specific facts and uses at issue in each case," the DOJ wrote, directly challenging a previous report from the U.S. Copyright Office that had rejected a blanket fair-use exemption for AI. "But it would be problematic—and legally incorrect—to impose broad copyright liability that would generally render training of AI models impermissible without licensing."[3]

Critics see a different motive behind the government's legal theory. The administration has reportedly discussed taking a 5 percent equity stake in OpenAI—a slice worth roughly $42.6 billion. While the brief declines to mention any prospective federal ownership, the alignment of interests has raised eyebrows among legal observers who view the national security framing as a convenient shield for corporate expansion.[4]

The government also argues that a strict licensing regime would paradoxically hurt smaller players. Requiring developers to pay for training data could create an oligopoly where only the wealthiest tech giants can afford to build foundational models, while simultaneously enriching legacy media organizations that hold massive, monetizable archives.[2][4]

A Statement of Interest does not bind U.S. District Judge Sidney Stein. But it signals a clear executive-branch posture heading into a critical period of litigation. With similar suits pending from authors, record labels, and other publishers, the DOJ has ensured that whenever a judge weighs the four factors of fair use, the specter of American geopolitical dominance will be sitting on the scale.[2]