DOJ AI Copyright Lawsuits Put National Security Before Fair Use

The Justice Department has entered the fight over AI training data with an unusually broad argument: courts should favor AI companies in more than a dozen copyright cases because U.S. “AI dominance” is a national-security interest. The DOJ’s statement of interest does not decide those cases, and judges are not required to follow it. But it puts the federal government squarely behind one side of a dispute that could shape creators’ control over their work. The DOJ filed its statement in early September.

What the government says, and what the filing supports

AI companies argue that training on copyrighted material is fair use because it transforms existing works into something new. The DOJ argued that rulings for creators would misunderstand fair-use doctrine, and pointed to potential uses such as intelligence analysis, improving weapons systems and battlefield recommendations. Those are arguments in the filing, not court findings.

The national-security case has a gap: the Atlantic’s account says the filing offered no evidence that large language models outperform Google search for the cited purposes, and no example of an LLM giving the military an important insight. The DOJ also argued that LLMs can help officials draw inferences from information in news articles. That connects AI to public information access, but it does not, by itself, show why copyright law should not apply to training on creative works. The reported filing makes both points.

A legal intervention, not a verdict

A statement of interest lets the executive branch advocate national goals in a lawsuit. It can influence a judge, but it cannot settle the legal question. The stakes are substantial: the Atlantic reports more than 130 lawsuits against OpenAI, Google, Amazon, Meta, xAI, Perplexity, Nvidia, Suno and other generative-AI developers. Plaintiffs including Encyclopedia Britannica, the publisher of Chicken Soup for the Soul, and adult-film company Strike 3 have sought payment for work they say was used without permission. Some plaintiffs seek damages in the billions of dollars.

The filing also matters because it may influence how courts approach the underlying cases without resolving them. University of Buffalo law professor Christine Bartholomew warned that frequent DOJ statements of interest could affect courts’ ability to develop independent responses and raise separation-of-powers concerns. Judges can reject the government’s recommendation, but the administration’s public pressure on courts may make that choice more complicated. The article describes both the concern and the limits of the filing.

The DOJ’s position could eventually affect how much control creators and publishers retain over their work. It does not establish that training is fair use, require creators to license their material, or settle what AI companies may owe. The consequential next step is judicial: rulings in cases involving OpenAI, Stability AI, Nvidia and others were expected in the coming months, and their outcomes could clarify where advocacy ends and copyright law begins. Those decisions had not yet been reported.