NYT Sues OpenAI While Joining 80-Publisher SPUR Coalition

Photo: A double octuple newspaper press built by R. Hoe & Co. of New York — the machine that industrialised the newspaper. A century later, publishers are fighting a different kind of machine. (Public domain illustration from Webster's New Illustrated Dictionary, 1911, via Wikimedia Commons)

The New York Times is suing OpenAI for copyright infringement — and this week it joined a publishers' coalition designed to stop companies like OpenAI from using its journalism for free. If that sounds like a contradiction, look closer: it is not. It is strategy.

On Monday, the Times, Agence France-Presse and The Economist signed on with more than 40 other publishers to the Standards for Publisher Usage Rights, or SPUR — taking the coalition past 80 members across more than 20 countries, according to Reuters. The new intake also includes The Atlantic and Brazil's National Association of Newspapers, joining founding members such as the Financial Times, the BBC and the Associated Press.

SPUR was launched in February and is funded entirely by its members. Its stated mission is to develop industry-wide standards and technical systems that let publishers track and control how AI companies use their content — and, crucially, to make licensing agreements easier to strike. In plain language: the people whose work trained the models are now organising against the people who built them.

Sue in the morning, standardise in the afternoon

The Times is running a deliberate two-track campaign. In court, its copyright case against OpenAI — with Microsoft added as a defendant this summer — grinds on in the consolidated proceedings in New York's Southern District, which reached the summary judgment stage in September. At the negotiating table, it has now taken a seat in the biggest publisher coalition the AI era has produced.

"Intellectual property protection and technological advancement can and must co-exist," Rebecca Grossman-Cohen, the Times' head of strategic partnerships and chief of staff, said in a statement. It is the kind of sentence designed to be quoted — and she is right that the two can co-exist. The open question is on whose terms, and at what price.

The logic of the double game is easy to follow. A lawsuit is a stick: slow, expensive, but capable of producing court orders and nine-figure settlements. A standards coalition is the bargaining table: faster, voluntary, and useful only if the other side shows up. The Times appears to have decided it needs both — the credible threat and the credible framework.

The $250 million warning shot

Four days before the SPUR announcement, the stick got considerably heavier. On 8 October, USA Today Co. — the publisher formerly known as Gannett — and 13 affiliated entities sued OpenAI in the US District Court for the Southern District of New York, alleging the company copied hundreds of thousands of their articles without permission to train the AI systems behind ChatGPT.

The 79-page complaint covers material published across 19 newspapers and news sites, including USA Today, The Tennessean, the IndyStar, The Columbus Dispatch and The Oklahoman, and names seven OpenAI entities as defendants. The publishers are seeking more than $250 million in damages, an injunction, and — the eye-catching remedy — the destruction of any AI models and training datasets built on their content.

The complaint's most awkward exhibit is numerical. Citing a Washington Post analysis, it says the plaintiffs' domains account for more than 122 million tokens in C4, a widely used training dataset — with about 83,266 entries from usatoday.com alone, according to further reporting on the filing. The suit also accuses OpenAI of reproducing their journalism in model outputs and of stripping out copyright management information such as bylines — a separate claim under the DMCA that carries its own statutory damages.

The case, numbered 1:26-cv-08892, is being run by Steven Lieberman of Rothwell, Figg, Ernst & Manbeck — the same lawyer acting for the Daily News plaintiffs, eight newspapers owned by Alden Global Capital including the Chicago Tribune, the Denver Post and the New York Daily News, in the consolidated copyright proceedings against OpenAI and Microsoft in the same court. OpenAI did not respond to requests for comment on the filing.

The pattern is unmistakable: local news, national news, wire services — the entire supply chain of American journalism is now suing the same defendants in the same courthouse. SPUR is what happens when the same industry decides to organise outside the courtroom as well.

Follow the 33 per cent

But training data is only half the grievance, and arguably the smaller half. The urgent one is traffic. According to a Reuters Institute report citing Chartbeat data, Google organic-search referrals to more than 2,500 websites fell 33 per cent globally between November 2024 and November 2025, as search engines rolled out AI-generated answers that let users consume information without ever clicking through to the site that produced it.

Read that number again and the coalition makes complete sense. Publishers are being charged twice: first their archives are ingested to train the models, then the models answer readers' questions directly and the referral traffic that pays for the next investigation never arrives. It is a colossally lopsided bargain — the raw material flows one way, the audience flows the other.

This is why SPUR matters more than yet another lawsuit. The lawsuits argue about the past: who took what, and what it was worth. The coalition is trying to price the future: every retrieval, every citation, every AI-generated answer built on a publisher's work, metered and reported back.

Telemetry: the union's actual weapon

The concrete idea at the heart of SPUR is the Content Telemetry Standard, which the coalition says it has recently introduced. It is an open-source technical specification that defines a common language for recording what AI systems do with publisher content — five categories: retrieved, grounded (used as the underlying evidence for an AI answer), cited, presented, and engaged with — and reporting that usage back to the content owner.

The telling detail, spotted by trade-press analysts who read the version 1.0 specification on the coalition's own repository, is the unit of measurement. The standard designates characters ingested and excerpt characters as the primary, portable counts, and demotes token counts to a supplementary role. A character is a character; a token is whatever a vendor's tokenizer says it is. Publishers did not write a manifesto. They wrote schemas with conformance tests — 171 test fixtures, 15 validated worked examples and 59 mutation checks run in continuous integration on every change. That is what makes the unit choice enforceable rather than aspirational.

A draft circulated in June, a public consultation ran through late July, and the 1.0 release extended the framework beyond text to images, video and audio. SPUR's technical lead, Alex Springer, has been blunt about the problem it solves: it is currently easier — and cheaper — for an AI firm to pull publisher content through free scraping credits than to sit down and negotiate a licence.

SPUR has invited OpenAI, Anthropic, Google, Meta and Microsoft to help shape the standard through a new invitation-only AI Licensing Advisory Board, expected to number around 20 members with a first meeting planned for October. So far, none of the AI companies has accepted. When Digiday asked Google whether it would join, a spokesperson answered only that the company "frequently engages with SPUR and other associations on a variety of topics" — the corporate equivalent of a polite shrug. The rest did not respond at all.

"This will ensure both sides of the equation have a seat at the table and can work together to drive through better, fairer and more sustainable ways of working," David Buttle, SPUR's co-founder, said in an emailed statement to Digiday. The sentiment is admirable. The empty chairs are the story.

Photo: A reconstructed early wooden hand press in a museum display — five centuries of printing technology built the industry now fighting AI firms over its content. (Photo: Andrew Plumb, CC BY-SA 2.0 via Wikimedia Commons; file photo)

Our take

Strip away the acronyms and SPUR is a publishers' union in all but name — and it is about time. For three years the industry's response to AI has been a shoddy patchwork: individual lawsuits, individual licensing deals, individual complaints about traffic, each publisher negotiating alone against companies worth hundreds of billions. Collective action is the obvious correction, and the telemetry standard is the first genuinely clever piece of it: you cannot charge for what you cannot measure, and until now nobody could measure anything.

But a standard nobody adopts is a ruler nobody agreed to be measured with. The AI labs' silence on the advisory board invitation tells you everything about where the power sits: they do not need SPUR, so they are in no hurry to legitimise it. That is precisely why the lawsuits matter — not as ends in themselves, but as the leverage that makes the ruler impossible to ignore. Whether the telemetry standard becomes enforcement infrastructure or an expensive dashboard showing publishers exactly how thoroughly they are being strip-mined will decide whether this coalition gets its members paid or merely gets them organised.

There is a precedent for optimism, and it is recent: Google has begun paying publishers for AI search content, a deal we covered last month, and Cloudflare has moved to rewrite the web's rules for AI crawlers at the infrastructure layer. Measurement, payment rails, crawling rules — the plumbing of a licensed AI web is being laid, piece by piece, by different actors. SPUR wants to own the metering layer. If even one major lab signs on, the rest will find the ruler very hard to refuse.

What to watch

Three things will tell you whether this coalition is real. First, the advisory board's October meeting: do any of OpenAI, Anthropic, Google, Meta or Microsoft actually take a seat? Second, the pilots SPUR says it is planning with technology and AI companies — the participants are unnamed, and names will matter more than press releases. Third, the courtroom: the consolidated New York cases are at summary judgment stage, and a decisive ruling on training-data fair use would reprice every licensing conversation overnight.

Before writing, we curled both of our earlier posts on this beat — on Google's publisher payments for AI search and on Cloudflare's attempt to rewrite the web's rules for AI crawlers — and confirmed each still returns HTTP 200. The context is live; the fight it describes is just getting started.

Sources

Reuters (12 October 2026) for the SPUR membership announcement and the Grossman-Cohen statement; Digiday (2 October 2026) for the Content Telemetry Standard 1.0 release, the advisory board invitations and the Buttle and Google statements; CFOtech and Media Copilot for the standard's technical detail; dig.watch, the Economic Times via Reuters, WKYU and Editor & Publisher for the USA Today Co. v. OpenAI filing of 8 October 2026. Figures on the referral collapse come from the Reuters Institute report citing Chartbeat data, via Reuters.

By TechPulse USA

TechPulse USA is an independent tech and AI news publication covering the stories shaping the industry — from Silicon Valley to London, Brussels, Toronto and Sydney. We verify every fact we print against multiple sources and say so when we cannot.