What happened
On October 30, 2026, Representatives Jerrold Nadler, Jamie Raskin, and Hank Johnson submitted an amicus brief in the Southern District of New York in *In re OpenAI, Inc. Copyright Infringement Litigation*. The announcement came from Nadler’s office, which describes the three members as ranking members of House Judiciary-related committees and subcommittees.
The brief opposes an argument advanced by the Department of Justice in support of AI companies. According to the lawmakers’ announcement, DOJ intervened on September 1 and argued that limiting the use of copyrighted text for AI development would endanger US AI development and national security. The lawmakers say the Copyright Act contains no national-security exception allowing private companies to use copyrighted works freely to train models.
Their central legal claim is narrow but consequential: national security is not one of the statutory fair-use factors. The announcement quotes the brief: “National security is not a fair use factor under the Copyright Act.” In the representatives’ account, treating it as one would add an exception that Congress did not enact.
The underlying dispute concerns allegations that OpenAI and Microsoft used New York Times journalism without permission to build AI products that compete with the newspaper. The announcement describes fair use as a doctrine permitting certain unlicensed uses after courts assess the circumstances: how a work was used, how much was copied, and whether the copying harms the market for the original. The lawmakers argue that DOJ is asking the court to bypass those statutes and cases in favor of a special rule for AI companies.
The brief asks the court to reject DOJ’s position and preserve legal protections that, in the lawmakers’ view, support compensation for writers, journalists, artists, and other creators. It does not resolve the case, establish the court’s view, or determine whether training on copyrighted material is fair use. It is an intervention in a live dispute over the legal conditions under which model developers may obtain and use training material.
What the archive adds
This is a new intervention in a known OpenAI copyright story. Our note, Fair use benefits everyone, covered OpenAI’s own fair-use argument in the New York Times case. The new fact is that House Judiciary Democratic ranking members are now contesting DOJ’s separate national-security rationale for the companies’ position.
Trend signal
No supplied New Runtime pattern is specific enough to characterize this court filing without overstating what it establishes.
Why now
The timing follows DOJ’s September 1 intervention supporting AI companies and comes as the New York litigation continues. The congressional filing is therefore a direct response to a particular argument already placed before the court, rather than a general statement on AI policy.
Connections to other events
The filing sits alongside a recent congressional effort directed at frontier AI labs: House Democratic committee leaders ask Anthropic and four other AI labs to report misalignment incidents within 48 hours. The topics differ, but both show House Democratic leaders using oversight mechanisms to press AI companies on issues they frame as requiring public accountability.
Why it matters
Teams building models, retrieval systems, and products around model APIs should distinguish a litigation filing from a legal rule. Nothing in this announcement changes a team’s permissions today. But the disagreement exposes a material uncertainty in the policy environment: whether arguments about US technological competition or national security can affect analysis of copyrighted training data.
That uncertainty matters beyond frontier-model developers. Vendors that supply data, build model layers, buy API capacity, or make content licensing decisions all operate under assumptions about what training inputs are available and on what terms. A court decision could affect those assumptions; this brief is evidence that the government’s pro-training rationale will be actively contested, including by members of Congress.
Limits and uncertainty
The public source is a statement from the filing lawmakers, not a court order or a neutral account of the parties’ positions. It presents their description of DOJ’s argument and of the consequences they believe it would have. The supplied material does not provide the full brief, DOJ’s complete filing, OpenAI’s or Microsoft’s response, a hearing date, or any indication of how the judge will rule. It also does not establish that national-security considerations can never arise elsewhere in copyright or AI policy; its claim is about their role as a fair-use factor in this litigation.
Skill to build now
Maintain a practical provenance record for datasets and externally supplied corpora: source, rights basis, acquisition date, permitted use, and downstream system that receives the material. This does not answer the fair-use question, but it gives a team a factual inventory if licensing requirements, vendor terms, or legal exposure change.
Reader takeaway
A group of House Judiciary Democratic leaders has asked the court not to let national-security arguments create a new path for unrestricted use of copyrighted journalism in AI training. The immediate operational change is zero, but the filing makes the legal and policy contest around model-training data more explicit.
Open question
Will the court address DOJ’s national-security framing in its fair-use analysis, or decide the case on the established statutory factors without reaching that argument?