Trump Administration Supports OpenAI in Copyright Lawsuit Against The New York Times

The Trump Administration has aligned with OpenAI in the artificial intelligence firm’s legal dispute with the New York Times.

In 2023, the media powerhouse filed a lawsuit against OpenAI and Microsoft, claiming that the companies breached copyright laws by training their AI systems on its articles without authorization. The government has now intervened, contending that OpenAI’s practices were permissible.

On Tuesday, the federal government sent a letter to the court asserting that the United States has a “strong interest” in determining whether training AI on copyrighted material constitutes “fair use,” as this impacts the nation’s ability to maintain “global leadership in artificial intelligence.” (The “fair use” doctrine allows specific uses of copyrighted content.)

The letter argued that the Times’ interpretation of fair use is “inconsistent” with established copyright laws and a favorable ruling for the newspaper could lead to severe repercussions. “Restricting LLM development based on a misinterpretation of fair use would stifle creative and scientific innovation while hindering American prosperity and economic mobility,” stated the government’s attorneys.

While the letter centers on the New York Times and OpenAI, the government clarified that its arguments are applicable to all parties involved in this case and related matters, including other publishers and authors.

The administration contended that training LLMs on copyrighted works is “extraordinarily transformative,” suggesting that these systems significantly alter their inputs. The officials also noted that OpenAI’s LLMs do not directly compete with The New York Times’ articles. Courts consider these factors, among others, when evaluating fair use claims.

Equating AI training with the outputs of artificial intelligence tools could create “problematic implications for copyright law as a whole,” implying it might be similarly unlawful for human writers to develop their skills using the works of others, according to the letter. (The lawyers drew a parallel to a young Joan Didion typing out Ernest Hemingway’s stories to refine her writing.)

Evan Brown, an intellectual property attorney at Neal & McDevitt, informs WIRED that US District Judge Sidney H. Stein overseeing this case is not required to consider the letter in his final ruling. However, he suggests that Stein and other judges in similar cases will likely “take it quite seriously because it originates from the Department of Justice and thus holds significant weight.”

Pamela Samuelson, a copyright expert co-directing the Berkeley Center for Law & Technology, tells WIRED that the government’s brief represents a “significant development” consistent with prior AI copyright rulings.

“The Administration is supporting a select few trillion-dollar AI firms at the expense of numerous American creators whose work they appropriated,” stated New York Times spokesperson Graham James to WIRED. “Both AI and creators can flourish—AI companies just need to compensate fairly for the content that enables their products, as copyright law mandates.”

OpenAI and Microsoft did not immediately respond to WIRED’s request for comments.

Other plaintiffs express dissatisfaction with the government’s involvement. Mary Rasenberger, CEO of the Author’s Guild—which filed its own lawsuit against OpenAI in 2023—told WIRED that the nonprofit is “extremely disappointed” with the letter, which she describes as “filled with flawed arguments and a gross misunderstanding of the fair use doctrine and copyright law.”

Numerous ongoing, high-profile lawsuits revolve around the legality of AI firms training their models on copyrighted works. The two most significant rulings so far occurred last year. Meta technically prevailed in its copyright case, Kadrey v. Meta, although the judge emphasized that the plaintiffs failed to provide sufficient evidence that the training had caused them harm while indicating that, under different circumstances, training on copyrighted materials without clearance might indeed be unlawful.