Seattle Times and Newsday Sue OpenAI and Microsoft in AI Copyright Battle

TL;DR

• The Seattle Times and Newsday have filed separate federal copyright lawsuits against OpenAI and Microsoft, alleging their articles were used without permission to train ChatGPT and Copilot.
• The publishers claim the AI models reproduce their journalism nearly verbatim, remove attribution, and undercut subscriptions by letting users bypass paywalls.
• The cases join a growing media-led legal battle that could decide whether AI training is fair use and force new licensing rules for news content.

What Happened

In back-to-back filings this week, two of America's most influential regional publishers took on the biggest names in AI.

The Seattle Times, Washington's largest daily, and Newsday, the Long Island-based daily owned by Cablevision founder Charles Dolan's Altice USA legacy group, each sued OpenAI and Microsoft in federal court, accusing the tech giants of building a multi-billion-dollar AI business on top of their journalism without consent or compensation.

Both complaints target OpenAI's GPT models and Microsoft's Copilot, which is built on OpenAI technology and integrated across Bing, Edge, Windows, and Microsoft 365. The publishers are seeking jury trials, statutory damages for alleged willful infringement, and court orders barring further use of their work in training and outputs.

The suits were filed separately but make strikingly similar arguments, and both law firms have previously worked with other publishers suing OpenAI.

What the Lawsuits Allege

At the core of both cases is a simple claim: OpenAI scraped millions of copyrighted news articles to train its large language models, then built products that compete directly with the publications that created them.

The Seattle Times complaint alleges that ChatGPT can recite large passages from its investigative reporting, local political coverage, and business journalism when prompted, sometimes word-for-word and with its paywall stripped away. Newsday makes a nearly identical claim, providing dozens of side-by-side examples where it says ChatGPT-4o and Copilot reproduced its Long Island, New York City, and national coverage with minimal prompting.

Both publishers bring counts for:

• Direct copyright infringement for ingesting and reproducing their articles in training datasets
• Removal of copyright management information, alleging OpenAI deliberately stripped author names, headlines, and copyright notices during training
• Unfair competition and unjust enrichment, arguing Microsoft and OpenAI free-rode on costly newsgathering to drive subscriptions to ChatGPT Plus and Copilot Pro
• Trademark concerns over hallucinations, where the AI allegedly fabricated stories and falsely attributed them to the papers, damaging reader trust

In its filing, Newsday said the practice is especially harmful for regional outlets that depend on local subscriptions and advertising. The Seattle Times echoed that, noting it employs more than 150 journalists whose work cannot be replaced by summaries.

Why These Publishers Joined the Fight Now

Neither paper is new to the AI debate, but until now neither had sued.

Executives for both outlets say they tried to negotiate. According to the complaints, OpenAI approached publishers over the past two years with licensing offers that the papers describe as taking it or leaving it - low fees, broad rights, and no guarantees against verbatim regurgitation or misattribution.

That frustration mirrors what drove The New York Times to sue OpenAI and Microsoft in December 2023, followed by suits from the Chicago Tribune, New York Daily News, Denver Post, Center for Investigative Reporting, and others owned by Alden Global Capital and MediaNews Group.

For The Seattle Times, which is independently owned by the Blethen family, and Newsday, which remains one of the largest suburban dailies in the U.S., the decision signals that the legal fight is no longer just national giants. Local news leaders argue they have the most to lose if AI search and chatbots answer questions using their reporting without sending readers - or revenue - back to the source.

Industry groups including the News/Media Alliance quickly praised the filings, calling unlicensed AI training one of the greatest threats to a sustainable press.

Part of a Growing Media Revolt

The two new cases land in an already crowded courtroom.

More than a dozen copyright suits against OpenAI, Microsoft, Meta, Google, Anthropic, and Perplexity are now winding through federal courts in New York and California. The New York Times case is furthest along and is widely seen as the bellwether, with a judge recently allowing key infringement claims to proceed while narrowing some unfair competition arguments.

At the same time, a parallel licensing market has emerged. The Associated Press, News Corp, Vox Media, The Atlantic, Dotdash Meredith, and the Financial Times have all signed paid deals with OpenAI, while others like Reddit, Stack Overflow, and People Inc. have licensed content to Google and others.

That split - sue versus sign - is now defining media strategy. Supporters of licensing say it provides vital revenue and citation traffic. Critics, including the plaintiffs in these new suits, say the fees are a fraction of the value extracted and do nothing to stop future scraping.

What OpenAI and Microsoft Are Saying

OpenAI and Microsoft have not yet filed formal responses to the Seattle Times and Newsday complaints, but both companies have used the same defense in prior cases.

OpenAI maintains that training AI models on publicly available internet content is protected fair use, similar to how a human learns from reading news. It argues its models transform content to create entirely new products, do not memorize articles except in rare edge cases it calls regurgitation bugs, and offers publishers an opt-out tool plus revenue-sharing partnerships.

Microsoft, which has invested more than $13 billion in OpenAI and integrates its models into Copilot and Bing, argues it is protected as a downstream technology provider and that publishers have failed to prove widespread, systematic copying by everyday users.

Both companies have previously asked judges to dismiss parts of the media lawsuits, with mixed success. Courts have so far refused to throw the cases out entirely, finding the examples of near-verbatim outputs plausible enough to go to discovery.

What This Could Mean for AI, Copyright, and the Future of News

Legal experts say the Seattle Times and Newsday suits add pressure in two critical ways.

First, numbers matter. Each new publisher adds thousands of additional registered copyrights to the potential damages pool. Under U.S. law, willful infringement can bring up to $150,000 per work, meaning liability could quickly reach into the billions if publishers prevail at trial.

Second, regional examples could resonate with juries and judges. While national investigations are high-profile, local scoops on city hall, schools, and public safety are harder for AI companies to claim are widely available elsewhere.

Three possible outcomes are now in focus:

A fair use ruling for AI: If courts side with OpenAI, it would cement broad rights to train on copyrighted web content, likely accelerating AI development but leaving publishers with little leverage beyond voluntary licensing.

A win for publishers: If training is found infringing, OpenAI and Microsoft could be forced to pay massive damages, delete models trained on news content, and build strict licensing and filtering systems - a precedent that would reshape the entire AI industry.

A mega-settlement and licensing system: Most analysts predict a middle path, where a ruling or trial pressure in the Times case forces a collective licensing framework similar to music royalties, with AI companies paying into a pool for news content.

For readers, the stakes are immediate. Microsoft and Google are already replacing traditional search links with AI answers. Publishers warn that without compensation or traffic, newsrooms will shrink, leaving less original reporting for AI to learn from in the first place - a scenario some researchers call model collapse.

Both cases are expected to be consolidated for pretrial proceedings alongside the other publisher suits in the Southern District of New York. Motions to dismiss are likely by late fall, with discovery extending well into 2027.

The message from Seattle and Long Island is clear: the AI copyright battle is no longer a one-paper fight. It is now an industry-wide showdown over who pays for the facts that power AI.

AndroGuider Team

Articles written by the AndroGuider team. We try to make them thorough and informational while being easy to read.