Trump Administration Backs OpenAI in Landmark New York Times Copyright Fight

The Trump administration has filed a statement of interest supporting OpenAI in the New York Times’ copyright lawsuit, arguing that training large language models on copyrighted text can qualify as fair use. The intervention matters because the case could shape how AI companies, publishers, and courts define the legal limits of model training in the United States.

The filing places the federal government on OpenAI’s side in one of the most closely watched copyright disputes in the AI industry. At stake is not only the Times’ claim that OpenAI and Microsoft unlawfully used its reporting to build AI systems, but also the broader question of whether AI development will continue to rely on large-scale ingestion of copyrighted material without direct licensing.

Why the government stepped into the case

The Justice Department is not a party to the lawsuit, but a statement of interest allows the government to tell the court how it believes the law should be interpreted. In this case, U.S. attorneys argued that a sweeping copyright rule against model training would conflict with core fair-use principles and slow technological progress.

In plain terms, the administration is telling the court that the law should not be read to block AI systems from learning from copyrighted text simply because the text is protected by copyright. The filing frames large language models as part of a broader innovation ecosystem, one that it says contributes to scientific research, productivity and economic mobility.

U.S. attorneys said the Times is trying to narrow fair use in a way that would exclude AI model training, warning that such a result would be inconsistent with copyright doctrine and would undermine the constitutional goal of promoting scientific and artistic progress.

The government’s position goes beyond this one lawsuit. It reflects a policy argument that has surfaced repeatedly in Washington as regulators, lawmakers and tech companies debate how to balance copyright enforcement with the rapid expansion of generative AI.

What the New York Times lawsuit alleges

The New York Times filed its lawsuit in December 2023, accusing OpenAI and Microsoft of using millions of its articles to train AI systems without permission or payment. The newspaper is seeking damages that it says could amount to billions of dollars, making the case one of the largest copyright battles ever brought against an AI company.

The Times argues that its reporting has commercial value and that OpenAI’s systems can reproduce, summarize or closely mimic its journalism in ways that threaten both its business model and the market for original news. The complaint also says the companies benefited from the newspaper’s work without compensating the organization that produced it.

OpenAI and Microsoft have pushed back, maintaining that model training is covered by fair use because the systems transform source material into statistical patterns rather than redistributing the material itself. That legal theory has become the central defense in many AI copyright cases.

How fair use is becoming the legal battleground

Fair use is the key legal question because it determines whether copyrighted material can be used without a license in certain contexts. Courts weigh factors such as the purpose of the use, the nature of the original work, how much was taken and whether the new use harms the market for the original.

AI companies argue that model training is transformative: the system is not republishing the articles, books or images it ingests, but learning from them to generate new outputs. Critics counter that the scale of data collection is unprecedented and that the commercial value of the underlying works is being exploited without consent.

The Trump administration’s filing aligns with the first view. Officials argued that imposing broad liability on training would be legally incorrect and would risk stalling a technology they see as strategically important to the United States.

What the statement says about American competitiveness

The filing links permissive training rules to national competitiveness, saying that constraining AI development would hinder American prosperity. That is a notable framing because it treats the copyright dispute not as a narrow industry fight, but as a question of whether the United States should keep its edge in AI.

Supporters of that view say innovation depends on access to large and diverse datasets, and that forcing developers to license every item individually would make model training impractical or prohibitively expensive. Opponents respond that ease of training should not override the rights of publishers, writers and artists whose work is used as input material.

This disagreement has turned copyright law into a proxy battle over the future structure of the AI economy.

Why this filing is politically and legally significant

The administration has increasingly used statements of interest in private cases to push policy goals, and one official has described the tactic as highly effective. By intervening in the Times case, the government is signaling that AI copyright doctrine is now a priority area for federal policy, not just a question for courts and industry lawyers.

The move is also politically notable because Donald Trump has an ongoing personal conflict with The New York Times, including a separate defamation lawsuit. That backdrop adds another layer to an already contentious case, though the Justice Department’s filing itself focuses on legal principles rather than the newspaper’s politics.

For OpenAI, the government’s support offers a powerful argument to cite as the company defends its training practices. For the Times, it means the lawsuit now carries not just corporate and media consequences, but a strong federal policy counterweight.

What this means for publishers and AI companies

The outcome of the lawsuit could influence how other media organizations choose between litigation and licensing. Some publishers have sued AI companies to protect their archives, while others have negotiated deals that turn their content into a revenue stream.

That split is already visible across the media landscape. News organizations that believe their archives are being used without compensation see lawsuits as a way to force bargaining power. Others have concluded that licensing agreements may offer a more practical path, especially if courts prove uncertain or slow.

The Times itself has taken both approaches over time. While it is suing OpenAI and Microsoft, it also struck a licensing arrangement with Amazon in 2025 allowing editorial content, including articles and recipes, to appear in Amazon’s generative AI tools. That dual strategy highlights how publishers are trying to protect value while still participating in the AI market when the economics make sense.

How other legal cases are shaping the field

Other lawsuits have begun to sketch the legal boundaries of AI training. A 2025 ruling involving Anthropic said the company could train models on books it had lawfully bought, but still faced liability tied to pirated material. That decision helped reinforce the idea that the source of the data matters, not just the fact that training occurred.

The same year, Anthropic reached a $1.5 billion settlement with authors, underscoring how expensive copyright disputes can become even when a company believes it has a defensible legal theory. Together, these cases suggest that courts may distinguish between licensed or lawfully obtained material and content acquired through piracy or other improper channels.

The Times case may go further, because it involves not just books, but journalism — a category of copyrighted work that publishers argue has direct market value and public significance.

What are the arguments on each side?

The core dispute is straightforward, even if the legal doctrine is not. OpenAI and the Trump administration argue that training is a transformative process protected by fair use, while the Times says its work was used commercially without consent and that the resulting AI systems threaten its market.

Below is a simplified comparison of the competing positions.

Why this case could become a precedent-setting fight

The Times lawsuit is being watched far beyond one newsroom. If the court sides with OpenAI on a broad fair-use theory, other AI companies may gain confidence that training on copyrighted text is lawful as long as the resulting system does not directly republish the source material. If the Times wins, the industry could face heavier pressure to negotiate licenses or alter training practices.

The stakes extend to every sector producing large amounts of protected content, including publishing, entertainment, education and software. A ruling that narrows fair use could force a major rethink of how foundation models are built. A ruling that expands it could accelerate AI development while intensifying criticism from creators and rights holders.

That tension explains why the case has attracted attention from lawyers, publishers, AI executives and policymakers. It is not just about a single newspaper or one AI lab. It is about whether the legal architecture that governs books, journalism and research can adapt to systems trained on vast libraries of digital material.

How the licensing trend is changing the debate

Licensing deals are becoming the compromise path in the market. At least several publishers have already reached agreements with OpenAI, including The Associated Press, Axel Springer and Vox Media. Those deals show that some rights holders would rather sell access than fight over whether training is allowed at all.

For AI firms, licensing can reduce legal exposure and improve relationships with publishers. For media companies, it can provide new revenue at a time when traditional advertising and subscription models remain under pressure. But licensing is not a complete solution, because not every publisher has the leverage to negotiate favorable terms, and not every AI developer is willing to pay for broad access.

That is why the legal system still matters so much. The existence of some licenses does not answer the bigger question of whether they are required.

Why this dispute goes beyond one company

The broader issue is whether AI systems should be treated like humans learning from the world or like commercial products that must clear rights before ingesting protected works. That philosophical divide sits underneath the technical and legal arguments.

Supporters of broad training rights say humans read books, articles and essays, then create new work based on what they have learned, so AI should be allowed to do something similar. Critics say the comparison fails because AI development is industrial, automated and built on a scale no individual reader could match.

That unresolved tension is why the Times case has become one of the defining disputes of the generative AI era.

What happens next?

The next steps will depend on how the court treats the government’s filing and how it assesses the copyright claims against OpenAI and Microsoft. The statement of interest does not decide the case, but it can shape the legal framing and influence the judge’s view of fair use.

Even if the lawsuit moves slowly, the policy signal is already clear: the administration believes AI model training should generally remain legally protected when it is tied to innovation and national competitiveness. That position is likely to feed into future litigation, lobbying and legislative debates well beyond this single courtroom fight.

For the media industry, the outcome could determine whether the future of AI and journalism is built on licenses, lawsuits, or some combination of both. For AI developers, it may decide how expensive and legally risky the next generation of models will be to train.

Either way, the New York Times case is now more than a copyright dispute. It has become a test of how the United States wants AI to grow, who gets paid when it does, and how far fair use can stretch in the age of large language models.