Times v. OpenAI Filings Turn on Copyright and Culture

Court filings submitted Friday in The New York Times's closely watched copyright lawsuit against OpenAI and Microsoft drew on a mix of copyright law and cultural references, with Microsoft's stock closing at 499.70, down 2.04%, and NYT at 67.28, up 0.33%.

The long-running copyright fight between The New York Times and the two biggest names in commercial artificial intelligence produced another round of paperwork on Friday, and the paperwork itself became the story. Filings entered in the newspaper's lawsuit against OpenAI and Microsoft reached beyond the dry mechanics of copyright doctrine, pulling in cultural and sporting references to make their points, according to NYT Technology.

That is not merely stylistic decoration. In a case with no close precedent, both sides are effectively arguing about what kind of thing a large language model is — a tool, a library, a student, a copyist — and analogy is the instrument they have to work with. When lawyers reach for culture and sport, they are usually trying to install a mental picture in a judge's head that the statute alone does not supply.

Why the analogies are doing legal work

American copyright law protects original expression, not facts or ideas. The fair-use doctrine — the exception that permits copying without permission in defined circumstances — asks courts to weigh the purpose of the use, the nature of the work, how much was taken, and the effect on the market for the original. Each of those four factors is a judgment call, and judgment calls invite comparison.

That is why the framing matters so much. If a model's training is cast as reading and learning, the fair-use argument gets easier. If it is cast as wholesale ingestion of a rival's product to build a substitute, the fourth factor — market harm — becomes the case's centre of gravity. The cultural references in Friday's filings are, in that light, a fight over which metaphor the court adopts before it ever reaches the arithmetic of damages.

Publishers watching from the sidelines have a direct stake. A ruling that leans toward the training-as-learning framing would weaken the leverage news organisations currently hold in licensing negotiations with AI developers. A ruling that emphasises substitution would strengthen it, and would likely reprice every content-licensing conversation in the industry.

How the two defendants sit differently

OpenAI and Microsoft are co-defendants, but they are not identically exposed. OpenAI built and trained the models at the centre of the dispute. Microsoft's involvement runs through its investment, its cloud infrastructure and the products it has shipped on top of the technology. Those are different kinds of alleged participation, and defendants in that position frequently end up with divergent legal strategies even while presenting a united front in public.

Microsoft is also the only one of the three parties that is both publicly listed and enormous. Its shares closed at 499.70 on Friday, down 2.04% from the prior close of 510.12, having traded between 499.36 and 511.00 through the session — a close at the very bottom of the day's range. That move sits against a mixed tape: the Nasdaq 100 tracker (QQQ) closed at $718.96, up 0.18%, while the S&P 500 tracker (SPY) ended at $770.19, down 0.39%, and the Dow tracker (DIA) at $534.08, down 0.53%. In other words, Microsoft underperformed a technology benchmark that finished the day higher.

No causal link between the filings and that move can be drawn from the available facts. Large-cap software trades on far more than litigation headlines, and single-session moves rarely encode legal risk cleanly. But the divergence is worth noting for readers tracking how markets are — or are not — pricing AI copyright exposure.

The plaintiff is a small-cap in a giant's fight

The New York Times Company's shares closed at 67.28, up 0.33% from a prior close of 67.06, with a session range of 66.39 to 68.00. The contrast in scale between plaintiff and defendant is one of the more striking features of this case: a newspaper publisher is litigating against a company whose market footprint dwarfs it, over the question of whether its archive can be used to build a product that may compete with it.

That asymmetry is part of why the case is so closely watched. The outcome will not be confined to one publisher. It will set the reference point for how every text-heavy business — news, reference publishing, academic presses, stock-image and archive libraries — negotiates with model developers. Licensing deals struck in the interim have generally been signed in the shadow of this litigation, with both sides aware that a court could shortly redefine what the underlying rights are worth.

What to watch as the docket moves

Friday's filings are a stage, not a verdict. The practical questions for anyone following the case from here:

• Whether the court accepts or rejects a fair-use framing at the summary stage. That determination shapes everything downstream, including whether a jury ever hears the case.
• Whether Microsoft's and OpenAI's defences begin to diverge. Co-defendants with different roles in the alleged conduct often separate as the record develops.
• How the market-harm factor is evidenced. This is the factor most amenable to hard proof, and the one most likely to turn on expert testimony rather than analogy.
• Whether a settlement emerges. Many disputes of this type resolve into a licensing arrangement, and a negotiated outcome would leave the underlying legal question unresolved for everyone else.

The wider bill for AI's training data

The economics behind the argument are straightforward even if the law is not. Frontier models require enormous volumes of high-quality text, and professionally edited journalism is among the highest-quality text that exists at scale. If that input has to be paid for at rates set by rights holders with legal leverage, the cost structure of training changes. If it does not, a large category of publishers loses a revenue line it had assumed was coming.

That is why the cultural flourishes in Friday's filings should not be dismissed as lawyerly showmanship. In a case where the statute was written long before the technology existed, the winning metaphor may be worth more than the winning citation. The filings landed at the close of a week in which the technology tape held up better than the broader market — a reminder that the legal risk hanging over AI development has, so far, been priced as a background condition rather than an immediate threat.

Frequently asked questions

What happened in the case on Friday?

Court filings were submitted in The New York Times's lawsuit against OpenAI and Microsoft. According to NYT Technology, the filings drew on a range of copyright law arguments alongside cultural and sporting references, using analogy to frame how artificial intelligence model training should be understood under existing copyright doctrine.

Why do cultural references appear in a copyright filing?

Fair use requires courts to weigh four open-ended factors, including the purpose of the copying and its effect on the market for the original. Because the statute predates modern AI, lawyers use analogies to give judges a workable mental model of what training a large language model actually resembles.

How are Microsoft and OpenAI positioned differently as defendants?

OpenAI developed and trained the models at issue. Microsoft's exposure runs through its investment, cloud infrastructure and the products built on the technology. Those are different forms of alleged involvement, and co-defendants in such positions often develop separate legal strategies as a case progresses.

How did the shares of the listed parties close?

Microsoft closed at 499.70, down 2.04% from a prior close of 510.12, in a day range of 499.36 to 511.00. The New York Times Company closed at 67.28, up 0.33% from 67.06, with a range of 66.39 to 68.00. Both figures are as of 20:00 GMT on September 4, 2026.

Did the filings move Microsoft's stock?

No causal link can be established from the available information. Microsoft closed down 2.04% while the Nasdaq 100 tracker QQQ finished up 0.18%, so the stock underperformed a technology benchmark, but large-cap software prices respond to many factors and single-session moves rarely isolate litigation risk.

Why does this case matter beyond one newspaper?

The outcome will set the reference point for how news publishers, reference and academic presses, and archive libraries negotiate licensing terms with AI developers. A ruling favouring fair use would weaken publishers' leverage; one emphasising market substitution would strengthen it and reprice content-licensing deals across the industry.

Sources

• Court Filings In A.I. Suit Invoke Copyright Law, Culture and Sports — NYT Technology

Photo: Phil Evenden · Pexels Licence — source