The US government doesn't usually weigh in on private copyright disputes between a newspaper and a tech company. On September 2, 2026, it did exactly that — and it sided against the newspaper.
What the DOJ actually filed
The Department of Justice filed a brief urging the court to rule in favour of OpenAI and Microsoft in the New York Times' copyright lawsuit over AI training data. This came after both sides had already filed cross-motions for summary judgment, following the Times' decision on June 25, 2026 to drop a contributory-infringement claim against OpenAI — narrowing the case's scope before the DOJ weighed in on what remained.
The bigger case this sits inside
The NYT suit isn't isolated — it's part of a consolidated multidistrict litigation bringing together 12 separate author and publisher lawsuits against OpenAI, combined since April 2025. The case is still in active discovery: a January 2026 court order already compelled OpenAI to produce 20 million anonymized ChatGPT logs, giving plaintiffs' counsel an unusually detailed window into how the system was actually used — discovery of that scale is itself a significant development in AI litigation, regardless of how the fair-use question eventually resolves.
Why a DOJ filing on the "OpenAI side" is significant
Executive-branch intervention in private copyright litigation is unusual, and it's not neutral — the DOJ's filing is an explicit signal about where the administration believes the fair use line should sit for AI training data. For publishers and content licensors, that's meaningfully different from watching two private parties litigate: it suggests the government's own policy preference leans toward permitting AI training on copyrighted material as fair use, at least in this case's specific posture.
What "fair use" actually turns on here
US copyright's fair-use doctrine weighs several factors — including the purpose and character of the use, and its effect on the market for the original work. The core question in cases like this is whether training an AI model on copyrighted news articles is a transformative use that falls within fair use, or whether it's a substitute that harms the market for the original journalism. The DOJ's filing doesn't resolve that question, but it puts real weight on one side of it.
Why this matters globally, not just in the US
Even though this is a US case under US copyright law, the outcome will shape how AI companies, publishers, and licensors worldwide think about AI training-data rights — a US ruling favourable to AI companies makes similar arguments more attractive for AI developers facing parallel disputes in other jurisdictions, while a ruling for publishers would strengthen the case for mandatory licensing regimes elsewhere.
Sources: Washington Post · Norton Rose Fulbright