Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Lawmakers Press Court to Reject DOJ's Fair-Use Arguments in NYT v. OpenAI

Three senior House Judiciary Democrats — Nadler, Raskin and Johnson — are urging the court to reject the Justice Department's argument that AI companies may train on copyrighted New York Times journalism freely and without limit. They frame the case as a test of whether copyright law still constrains how models are built.

Published

In a statement issued by Rep. Jerrold Nadler's office, Reps. Nadler, Jamie Raskin and Hank Johnson pressed the court to reject the Justice Department's position in the copyright litigation over AI training data.

The substance of their argument is narrow but consequential: DOJ has argued that AI companies should be able to use the Times' copyrighted journalism to train models without payment or meaningful restriction, and the lawmakers say the court should not adopt that view.

The underlying dispute is whether ingesting news archives and using them to train commercial models counts as fair use. Courts weigh factors including the purpose of the use, the nature of the work, how much was taken, and the effect on the market for the original.

Fair use has historically been a bounded exception for commentary, teaching and news reporting. Stretching it to cover large-scale commercial training effectively asks judges, rather than Congress, to set the data rules of the AI era.

In a parallel public push, OpenAI published a post compiling excerpts from court briefs by creators, academics, accessibility advocates and legal experts arguing that fair use supports AI training and innovation — a reminder that this fight is now conducted both inside and outside the courtroom.

For AI developers, the stakes are structural. If training data must be licensed item by item, the marginal cost of building frontier models rises sharply; if fair use is read broadly, the bargaining power of publishers weakens.

For news organizations, the concern goes beyond licensing revenue. A model that absorbs reporting and answers readers directly competes for the same attention, and copyright protection for factual works is already thin.

Watch three things: how much weight the court gives DOJ's position, whether other pending AI copyright suits adopt the same reasoning, and whether Congress moves to legislate on training data. The lawmakers' choice to file rather than legislate suggests the statutory route is not moving quickly.

Why it matters

The outcome will shape whether frontier models must license their training data, affecting both AI development costs and publishers' leverage — and with Congress moving slowly, the rule may be set by judges first.

CopyrightPolicyOpenAI
Back to realtime news

Nearby Updates

All