Justice Department Supports OpenAI's Fair Use Defense in Copyright Dispute
The Justice Department has recently filed a statement of interest backing OpenAI's fair use defense in a series of consolidated copyright cases involving publishers. The department's argument centers on the assertion that high licensing costs could threaten national security by making it more difficult to develop a robust American AI industry. According to the filing, such regulations would inadvertently provide a competitive advantage to foreign adversaries who are not subject to the same constraints.
The legal proceedings are taking place in a New York court, where the U.S. government has stated that training AI models on publishers' content constitutes fair use. This position was reported by CNET on Tuesday. The Justice Department's filing emphasizes that the current rules could hinder the growth of the AI sector in the U.S., which is critical for maintaining national security.
In contrast, The New York Times has publicly rejected this argument. A spokesperson for the publication, Graham James, criticized the administration for siding with large AI companies, asserting that these companies should fairly compensate content creators for their work.
The cases are being heard in the Southern District of New York, with notable lawsuits filed by The Times in December 2023 and Ziff Davis, the parent company of CNET, in 2025. Additionally, around 400 local newspapers have also initiated legal action.
In Europe, the situation is markedly different. EU copyright law does not recognize a fair use doctrine, instead offering a limited set of exceptions. The EU's text and data mining exception allows rightsholders to opt-out under Article 4(3) of the 2019 copyright directive, which complicates the legal landscape for AI model providers.
The AI Act in the EU directly relates to these copyright obligations, requiring that any general-purpose model provider must identify and respect the rights reserved by content creators. This requirement applies regardless of where the training of the AI model occurred.
The Justice Department's argument appears to be a response to these European regulations, which aim to prevent any provider from gaining an unfair advantage in the EU market by adhering to lower copyright standards. A recent ruling in Munich found against OpenAI, determining that lyrics memorized by the GPT-4 model constituted reproduction, a decision that is currently under appeal and may reach the Court of Justice.
As the legal battles unfold, it remains to be seen how the outcomes in the U.S. will impact the obligations of AI companies operating in Europe, where publishers have already raised concerns about the withholding of evidence in these cases.
