The Trump administration just drew a sharp line in one of the most contentious battles in technology. On Tuesday, the Justice Department filed a detailed brief in Manhattan federal court backing OpenAI against The New York Times and a group of other publishers. Training large language models on copyrighted material counts as fair use, government lawyers argued. Restricting it would damage American scientific progress, economic strength, and national security.
Short and direct. The filing marks the first time federal officials have taken a formal position on the merits in the wave of copyright suits targeting AI companies. It lands amid growing tension between creators who see their work absorbed without payment and tech giants racing to build ever more powerful systems.
And the stakes could hardly run higher. Large language models sit at the heart of tools millions use daily. They summarize articles, generate code, answer questions, and increasingly influence decisions in business, medicine, and defense. Without vast troves of high-quality text to study, those models lose coherence and capability. Or so the argument goes.
The Wall Street Journal first reported the government’s intervention. Associate Attorney General Stanley E. Woodward Jr. and Assistant Attorney General Brett Shumate signed the document. They told the court that limiting AI companies’ access to published works would prove “inconsistent with basic copyright law principles.” It would also “thwart creative and scientific progress.”
But not everyone sees progress. The New York Times sued OpenAI and Microsoft in late 2023. The paper accused the companies of feeding millions of its articles into training data without permission or compensation. Outputs from ChatGPT sometimes reproduced near-verbatim excerpts, the suit claimed. That competition directly harms the newspaper’s subscription business.
Graham James, a spokesman for the Times, pushed back hard. “The administration is siding with a handful of trillion-dollar AI companies at the expense of the countless American creators whose work they stole,” he said, according to The New York Times. “The administration’s proposal to let companies take that content without permission or compensation would undermine the sustainability of the human-created content that a healthy society depends on, and which AI needs to function.”
Strong words. They reflect a broader frustration among authors, journalists, book publishers, and musicians. Many have filed parallel suits. Some AI developers settled. Anthropic, for instance, agreed to pay $1.5 billion to resolve claims it used pirated books. Others fight on.
The government brief takes a different view. It calls training “extraordinarily transformative.” The models do not store or regurgitate exact copies in most cases. Instead they learn statistical patterns across enormous datasets. That process creates something new. Something useful. Something different from the original texts.
Courts weigh four fair-use factors. The purpose and character of the use. The nature of the copyrighted work. The amount taken. And the effect on the potential market. Government lawyers emphasized the first and fourth. Training serves a research-like purpose. And ChatGPT does not substitute for reading the Times, they said. It does not replace the original articles in the marketplace.
They went further. Constraining development “would thwart such creative and scientific progress while hindering American prosperity and economic mobility.” National security entered the picture too. Advanced AI already aids intelligence analysis and weapons systems, according to a Government Accountability Office report cited in the filing. Rules that hobble U.S. firms would hand advantages to foreign adversaries less constrained by copyright concerns.
The message echoes statements made the same day at a G20 gathering in Chapel Hill, North Carolina. Commerce Secretary Howard Lutnick urged other nations to embrace fair use for AI training while finding ways to protect artists. “We need frameworks that protect creators and inventors whose ingenuity drives our economies without creating the barriers that prevent the next generation of innovators,” he told officials, Reuters reported.
Tech executives joined the chorus. Nvidia’s Jensen Huang, OpenAI’s Sam Altman, Meta’s Mark Zuckerberg, and Elon Musk warned against heavy regulation. Innovation moves faster than rules, they suggested. The administration appears to agree.
This stance builds on an executive order President Trump signed earlier in 2025. It aimed to sustain U.S. leadership in AI and promote economic competitiveness. The brief references that order directly. A narrower reading of fair use would hamper it, officials wrote.
Yet the AppleInsider coverage captured an underlying anxiety in Washington. Government officials worry AI companies genuinely cannot maintain their pace of innovation without access to broad copyrighted material. Call it legal theft or transformative learning. The practical effect remains the same. High-quality English-language books, news archives, and creative writing provide the richest fuel for models that must understand nuance, context, and human expression.
Licensing everything at scale presents massive problems. Only the largest players could afford comprehensive deals. Smaller innovators would struggle. That concentration risks reducing competition rather than spurring it. The brief explicitly warns against creating an oligopoly through licensing barriers that function as subsidies for established media companies.
Critics counter that creators deserve payment. Their work built the internet’s knowledge base. AI systems now profit from it. Without some revenue stream back to journalists, novelists, and photographers, the quality and quantity of new material could decline. AI needs fresh human content to stay relevant. A vicious cycle emerges if incentives collapse.
The filing does not resolve every question. It addresses training specifically. Outputs that reproduce protected expression could still trigger liability. Memorization and regurgitation remain live issues in discovery disputes. Judges must still weigh evidence case by case.
Even so, the intervention carries weight. Courts often give deference to statements of interest from the United States. This one signals clear policy direction from the highest levels. It could influence judges in the consolidated OpenAI copyright litigation before U.S. District Judge Sidney Stein. Similar briefs may appear in other suits.
International dimensions complicate matters. The European Union takes a stricter approach. Its AI Act and copyright directives require more transparency and offer opt-out mechanisms for text and data mining. Companies selling models in Europe must navigate those rules regardless of where training occurred. China operates with fewer apparent constraints, though exact practices stay opaque.
That global patchwork explains part of the urgency. U.S. officials see AI leadership as a strategic imperative. They fear falling behind if domestic rules prove too burdensome. The brief mentions foreign adversaries twice but avoids naming China directly. The subtext feels clear enough.
Publishers and creators will not surrender quietly. They point to existing licensing markets. News organizations already sell content to AI firms under commercial agreements. Some deals have been struck quietly. The Times itself has negotiated with certain players. Why not require payment across the board?
The government answers that transaction costs would prove prohibitive for the scale involved. Billions of documents. Trillions of tokens. Individual negotiations could slow development to a crawl. And many older works sit in the public domain or under licenses that permit such uses.
So the debate continues. Fair use has always balanced incentives for creation against the spread of knowledge. Congress wrote the doctrine with flexibility in mind. Courts have stretched it to cover search engines, parody, and software reverse engineering. Now they face its application to machine learning at unprecedented scale.
Woodward called the filing “a historic statement of interest.” On social media he added that President Trump made clear AI dominance matters for national security, prosperity, and economic mobility. The administration has chosen its priority. For now, that priority tilts toward rapid AI advancement.
Whether that choice holds as cases climb the appeals ladder remains uncertain. The Supreme Court may eventually weigh in. Congress could legislate a clearer framework. For the moment, the Justice Department’s brief tilts the scales.
Tech companies cheered quietly. Publishers voiced outrage. Creators watched with concern. And the models kept training. On mountains of text. Some paid for. Much of it not. The race accelerates. The legal fight follows behind.
One thing feels certain. The tension between protecting intellectual property and unlocking artificial intelligence will define technology policy for years. Wednesday’s filing did not settle the argument. It simply made the government’s bet explicit.
Washington Picks Sides in AI Copyright War: Innovation or Theft? first appeared on Web and IT News.