Justice Department Backs OpenAI in Times Copyright Case
Washington has taken a side in the highest-profile AI copyright fight, telling a court that OpenAI's training on Times journalism was lawful and tying the outcome to national security.
The U.S. Justice Department filed a court brief on September 2, 2026 arguing that OpenAI's use of New York Times articles to train artificial intelligence models was lawful, and that American dominance in AI is a national security interest.
The federal government has stepped into the most closely watched copyright fight of the artificial intelligence era, and it has stepped in on the side of the technology company. In a court filing, the Justice Department argued that OpenAI acted lawfully when it used articles published by The New York Times to develop its A.I. systems, and it framed the question as one bearing on national security, telling the court that American dominance in technology depends on the outcome.
That is an unusual posture. Copyright disputes between a publisher and a technology firm are ordinarily private litigation, fought over the four statutory fair use factors and settled or tried without the executive branch weighing in on who should win. The department's decision to file at all signals that Washington now views the legality of A.I. training data as an industrial policy question rather than a purely commercial one.
What the department actually argued
Two claims sit at the center of the filing, according to NYT Technology. The first is legal: that OpenAI's use of Times articles to build its models was permitted under existing law. The second is strategic: that national security depends on the United States staying ahead in technology.
The pairing matters more than either piece alone. Fair use is a fact-bound, four-factor test — the purpose of the use, the nature of the work, how much was taken, and the effect on the market for the original. Nothing in that framework asks a judge to consider geopolitical competition. By attaching a national security argument to a fair use position, the government is inviting courts to read the doctrine with an eye toward where the technology industry ends up, not only toward what happened to a particular set of articles.
Publishers will read that as an attempt to change the subject. Their central claim in these cases has always been the fourth factor: that models trained on their archives can substitute for the journalism itself, damaging subscription and licensing markets that news organizations spent decades building. A brief about American competitiveness does not answer that. It reframes the stakes so that a ruling for the publisher looks like a cost imposed on the country.
Why one brief can move an entire docket
The Times case is not isolated. It sits at the front of a long queue of suits brought by authors, artists, music publishers, image libraries and news organizations, all asking the same underlying question: does ingesting copyrighted work to train a model require a license, or is it transformative use that does not?
Courts have been working toward an answer case by case, and the reasoning in the earliest decisions tends to get cited in the later ones. A statement of interest from the Justice Department does not bind any judge. But it is the kind of document that gets quoted in every subsequent brief on the defense side, and it gives a court that is inclined to rule for the developers a well-credentialed argument to lean on.
It also changes the negotiating table. Several A.I. developers have signed content licensing agreements with publishers rather than test the legal question. Every one of those deals was priced against the risk of losing in court. If the federal government's position makes a defense verdict look more likely, the price a publisher can command for its archive falls. That is a real economic consequence for the news industry regardless of how the Times case itself ends.
How the market read it
The immediate market reaction was muted, which is what you would expect from a filing that changes the odds in a case rather than resolving it. Shares of The New York Times Company (NYT) traded at 68.29, up 0.83% on the day, within a session range of 67.44 to 68.53 and above the prior close of 67.73, as of 18:46 GMT on September 2, 2026. Microsoft, the largest listed company with direct commercial exposure to OpenAI, traded at 495.20, down 1.16% from a prior close of 501.02, with a day range of 493.81 to 500.27 at the same timestamp.
Neither move is a verdict on the litigation. The broad tape was quietly higher: the S&P 500 tracker (SPY) at $764.82, up 0.40%; the Dow 30 tracker (DIA) at $529.83, up 0.39%; and the Nasdaq 100 tracker (QQQ) at $708.09, up 0.06%. In other words, the Times outperformed a modestly positive market and Microsoft underperformed a flat one — a spread far too small to attribute to a court filing with any confidence.
That muted response is itself informative. Investors have long since concluded that copyright litigation is a cost line for the large model developers rather than an existential threat, and that for a publisher the value at stake is licensing revenue rather than the core business. A single brief does not disturb either assumption.
The questions the filing leaves open
Several things remain unresolved, and they are what to watch from here.
- Whether the court accepts the framing. A judge can acknowledge the government's interest and still decide the fair use question strictly on the record before it. Courts are generally resistant to importing policy arguments into statutory tests.
- Whether other agencies or states line up differently. An executive branch position on A.I. training does not preempt state consumer protection or contract claims, and it does not bind foreign regulators, several of whom have taken a markedly more protective line on text and data mining.
- What happens to licensing deals already signed. Contracts negotiated in the shadow of legal uncertainty do not unwind because the uncertainty shifts, but renewals will be priced differently.
- Whether this becomes routine. If the department files in one A.I. copyright case, publishers in the others will ask why it did not file in theirs — and developers will ask that it does.
What is really being decided
Strip away the procedural detail and the case is about who captures the value in the chain that runs from a reporter's work to a chatbot's answer. Publishers argue that the archive is the input and should be paid for as one. Developers argue that learning from text is not copying it, and that no one has ever needed a license to read.
Until now, that argument has been conducted between private parties in front of judges. The Justice Department's filing puts a third voice in the room and attaches the phrase national security to a fair use brief. Whatever the court decides, the terms of the debate have changed: the question is no longer only what the law permits, but what the government wants the answer to be.
Key facts
- Filing: Justice Department brief supporting OpenAI in The New York Times copyright suit, September 2, 2026
- Core argument: OpenAI's use of Times articles for A.I. training was lawful; U.S. tech dominance framed as a national security interest
- NYT share price: 68.29, +0.83%, as of 18:46 GMT on September 2, 2026
- MSFT share price: 495.20, -1.16%, as of 18:46 GMT on September 2, 2026
Frequently asked questions
What did the Justice Department argue in its filing?
The department told the court that OpenAI legally used articles published by The New York Times to develop its artificial intelligence systems. It went further than the copyright question alone, adding that national security depends on the United States maintaining dominance in technology — an argument that ties the case's outcome to American competitiveness rather than to the parties' commercial dispute.
Does a Justice Department brief decide the case?
No. A filing from the federal government does not bind the court and does not resolve any claim. It is persuasive material that a judge may weigh, ignore, or address directly. Its practical influence comes from how often it will be cited in later briefs and from how it shifts perceived odds among parties negotiating licensing deals.
What is fair use in a copyright case?
Fair use is a defense in U.S. copyright law that permits unlicensed use of protected work in certain circumstances. Courts weigh four factors: the purpose and character of the use, the nature of the copyrighted work, how much of it was used, and the effect on the market for the original. It is decided case by case on the specific facts.
How did the shares react?
The reaction was small. The New York Times Company traded at 68.29, up 0.83% on the day, as of 18:46 GMT on September 2, 2026. Microsoft traded at 495.20, down 1.16%. Both moves are within ordinary daily ranges and cannot be reliably attributed to the court filing alone.
Why does this matter beyond the Times case?
Numerous authors, artists, image libraries and publishers have filed similar suits over A.I. training data. Early rulings shape later ones, so a federal position favoring developers strengthens the defense across the whole docket. It also affects pricing power in content licensing negotiations, since those deals are priced against the risk of losing in court.
Is OpenAI a publicly traded company?
No. OpenAI is not listed on a public exchange, so investors cannot buy its shares directly. The nearest listed exposure is Microsoft, which has a commercial relationship with the company and traded at 495.20, down 1.16%, as of 18:46 GMT on September 2, 2026, according to licensed market data.
Sources
Photo: Ramon Perucho · Pexels Licence — source
