WASHINGTON — The Trump administration has sided with OpenAI in its copyright battle with The New York Times and other publishers, arguing that training artificial-intelligence models on copyrighted material can qualify as fair use and that restricting the practice could damage US technological and economic interests.
The Justice Department filed a brief in Manhattan federal court on Tuesday, marking the first known intervention by the US government in the growing number of lawsuits brought by copyright holders over the use of their work to train AI systems. The filing does not carry the force of a court ruling, but gives OpenAI and other technology companies backing from the federal government as they defend their model-training practices.
The government said the United States has a strong interest in rejecting arguments that training large language models on copyrighted texts necessarily violates copyright law. It pointed to scientific research, national security, economic development and the country’s competition with foreign rivals as reasons for allowing the technology to develop.
Associate Attorney General Stanley Woodward said the administration would not allow what it considers an incorrect interpretation of copyright law to leave the United States at a disadvantage to foreign competitors.
“AI dominance is critical to promote national security, prosperity, and economic mobility for all Americans,” Woodward said in a statement posted on X.
The dispute centres on whether copying copyrighted works to train AI models constitutes fair use, a legal doctrine that permits certain unlicensed uses of copyrighted material. The government argued that AI training is highly transformative because the systems process large volumes of material to develop models capable of generating new content.
The Justice Department said the creative and public benefits of training large language models outweigh potential competitive harm to copyright owners. It also argued that limiting access to copyrighted material for AI training could impede scientific and technological advances.
The New York Times has taken the opposite position. It sued OpenAI and Microsoft in 2023, accusing the companies of using millions of its articles without permission to develop and operate AI products, including ChatGPT.
The Times has argued that allowing AI companies to use publishers’ work without payment threatens the economic foundation of journalism and other creative industries. A spokesperson for the newspaper said the administration was siding with large technology companies rather than creators whose work was used to develop AI systems.
“Both AI and creators can thrive — AI companies simply need to pay fairly for the content that makes their products possible, as copyright law requires,” the spokesperson said.
OpenAI and Microsoft have denied the allegations and maintain that their use of copyrighted material is protected by fair-use principles. OpenAI did not immediately respond to requests for comment on the government’s filing.
The Times case is part of a much broader legal fight. Authors, publishers, musicians and other copyright holders have filed lawsuits against AI companies including OpenAI, Anthropic and Meta Platforms, alleging that copyrighted material was used without authorisation to train generative AI systems.
The cases could help establish how US copyright law applies to the rapidly developing AI industry. At the centre of many of them is the question of whether an AI model’s use of copyrighted works to learn patterns and generate new material is sufficiently transformative to qualify as fair use.
Courts have so far produced mixed signals. The first two judges to consider closely related AI copyright questions issued differing rulings last year, leaving companies and content owners without a settled legal standard.
The administration is also seeking support abroad for its approach. US Commerce Secretary Howard Lutnick told officials from G20 countries at a meeting in North Carolina on Wednesday that governments should embrace fair-use principles that permit AI companies to train models on creators’ work while developing safeguards to protect artists.
The government’s intervention comes as the Trump administration has made maintaining US leadership in artificial intelligence a national priority. Its position could carry implications beyond the Times case because a ruling on AI training and fair use is likely to affect the wider set of lawsuits challenging how technology companies obtain and use material for model development.
The federal court will ultimately decide whether OpenAI’s use of copyrighted works falls within the protection of US fair-use law. Until then, the government’s filing gives the AI industry a powerful new argument: that the legal treatment of model training is not only a question of copyright, but also one of national competitiveness and technological policy.
