Justice Department sides with OpenAI in publishers’ copyright fight

The Justice Department filed a statement of interest backing OpenAI’s fair use defence in the consolidated publisher copyright cases, arguing that licensing costs would threaten national security. EU law has no fair use doctrine, and the AI Act applies its copyright obligations to any provider placing a model on the Union market wherever training occurred.

The US government has told a New York court that training AI models on publishers’ work is fair use. The Justice Department filed on Tuesday, CNET reported.

The argument offered is national security. Rules that make a robust American AI industry harder to build “threaten national security and give a competitive advantage to foreign adversaries who are not so encumbered“, the filing says.

It goes further. The statement of interest also argues that licensing costs would hand the largest technology companies an oligopoly on model training.

The New York Times rejects that. Spokesperson Graham James said the administration is siding with a handful of trillion-dollar AI companies, and that AI companies simply need to pay fairly for the content that makes their products possible.

The cases are consolidated in the Southern District of New York. The Times sued in December 2023, Ziff Davis, which owns CNET, sued in 2025, and 400 local newspapers have their own action.

None of this reasoning is available in Europe. EU copyright law has no fair use doctrine, only a closed list of exceptions.

What Europe has instead is a text and data mining exception with an opt-out. Rightsholders can reserve their rights under Article 4(3) of the 2019 copyright directive, and the exception then stops applying.

The AI Act plugs straight into that mechanism. Article 53 requires every general-purpose model provider to operate a policy identifying and respecting those reservations.

It attaches to the model rather than to the training run. Recital 106 says any provider placing a model on the Union market must comply regardless of the jurisdiction in which the copyright-relevant acts took place.

The recital states its own purpose. No provider should be able to gain a competitive advantage in the Union market by applying lower copyright standards than Europe’s.

That is the Justice Department’s argument written backwards. A Munich court ruled against OpenAI in November, finding that lyrics memorised in GPT-4 amounted to reproduction and that the mining exception did not cover it.

That judgment is under appeal and may reach the Court of Justice. It was reached on conduct the Justice Department now says is protected.

So a win in Manhattan would not travel. OpenAI would face the same European obligations in the market where publishers have already accused it of withholding evidence.