Source: R Street Institute

In a statement of interest filed in the New York Times’ lawsuit against OpenAI, the Department of Justice urged the court to reject the Times’ argument that training AI models on copyrighted works is itself copyright infringement and to treat training as a transformative use distinct from downstream model outputs. The Times sued OpenAI in December 2023, alleging that training on its articles without permission infringed copyrights; a federal judge allowed core claims to proceed and plaintiffs seek remedies that include destruction of the offending models and training data. The government contends that creating a broad copyright requirement for training would hamper competition among AI developers and effectively subsidize the largest publishers by directing licensing revenue to legacy outlets.

The filing frames the dispute in terms of local journalism’s vulnerability: about 212 U.S. counties lack a locally based news source and another 1,525 have only one, while the Times has nearly 13 million subscribers and competes with local outlets for limited subscription dollars. The DOJ warns licensing fees would disproportionately benefit high-volume legacy publishers and erect entry barriers that harm small newsrooms. The commentary highlights philanthropic and programmatic uses of AI for local reporting — including a $10 million Lenfest Institute program and grants for AI-powered public-records tools and newsroom chatbots — and warns that a licensing mandate or destruction remedy could raise costs, narrow capabilities, and remove tools already used by local newsrooms. The piece urges courts to decide fair use on legal grounds while assessing any rule by its effect on the smallest newsrooms rather than the largest.