The New York Times, the Daily News, and other major news organizations are demanding that a federal judge impose sanctions on OpenAI for allegedly hiding critical evidence in what could become a landmark copyright case over artificial intelligence and journalism.
Thursday's filing in Manhattan federal court accuses the ChatGPT maker of deliberately obstructing the legal process by refusing to release datasets and system logs that would show how OpenAI and Microsoft built their AI technologies using millions of news articles without permission or payment. The newspapers argue that OpenAI "chose obstruction" over transparency, and that recent depositions reveal the company has been misleading the court about its ability to search for copyrighted content in its training materials.
"This motion asks the court to punish OpenAI for hiding and destroying evidence showing how ChatGPT was trained on stolen journalism," said Steven Lieberman, attorney for the New York Daily News and seven sister publications. The legal team is seeking attorney fees and other penalties for what they characterize as two years of "misrepresentations" about the company's data practices.
The Core Question: Who Pays for News?
At the heart of this dispute is a fundamental question about the future of journalism: Should artificial intelligence companies be allowed to train their systems on the work of professional journalists without compensation or consent? The newspapers argue that AI chatbots are now competing directly with news organizations as information sources, siphoning off web traffic that once generated advertising revenue—the lifeblood of newsrooms.
The threat became starkly visible about two years ago when Google introduced AI-generated summaries at the top of online search results, cutting off the advertising dollars that flow when readers click through to the original reporting. That shift underscored what news organizations have long warned: AI companies are building profitable products on the backs of journalists' work without sharing the gains.
The New York Times alone has spent more than $28 million fighting AI companies in court, according to financial disclosures. That figure includes a separate lawsuit the newspaper filed against AI company Perplexity. For an industry already struggling with declining print revenues and digital competition, these legal costs represent a massive drain on resources that could otherwise fund newsroom operations.
OpenAI's Defense and the Fair Use Question
OpenAI and Microsoft, the company's business partner, have defended their practices under the "fair use" doctrine of copyright law, arguing that training AI systems on publicly available internet content falls within legal bounds. The company also contends that withholding detailed ChatGPT logs is necessary to protect user privacy.
"As the Times' case weakens and they've been forced to drop claims against us, they're persisting with their efforts to invade the privacy of people who have nothing to do with this case," OpenAI spokesperson Drew Pusateri said Thursday. "We'll continue defending our users' privacy and the long-established principles of fair use."
But the newspapers counter that privacy concerns don't justify hiding evidence about how the AI system was trained. The legal dispute reflects a broader tension in the AI industry: whether companies can claim privacy protections while simultaneously refusing to disclose how they built systems on copyrighted material.
A Pattern Across Creative Industries
The copyright battle over AI isn't limited to journalism. Visual artists, novelists, music record labels, and other creative industries have filed dozens of lawsuits against AI companies, with mixed results. In the largest copyright settlement so far, OpenAI rival Anthropic agreed to pay book authors $1.5 billion for training its Claude chatbot on pirated works—a sum that represents only a fraction of Anthropic's $965 billion market valuation as it prepares to go public.
That settlement suggests significant financial liability exists for AI companies that use copyrighted material without permission. Yet it also reveals how small such payments can be relative to the companies' overall valuations, raising questions about whether legal penalties will meaningfully change corporate behavior.
The Licensing Alternative
Meanwhile, some news organizations have begun negotiating licensing deals with AI companies, including OpenAI, Google, and Meta. These arrangements typically involve the news outlet receiving a fee in exchange for allowing the company to train AI systems on their content. The Associated Press announced such a deal with OpenAI about three years ago, establishing a model that other outlets are now exploring.
These agreements represent a middle path between litigation and the status quo—one where news organizations receive compensation for their work. But they also raise concerns about which outlets can afford to negotiate favorable terms and which smaller publications might be left behind without any compensation as AI companies continue training their systems.
Why This Matters:
This case will likely determine whether AI companies must compensate journalists and news organizations for the content that trains their systems, or whether they can continue using published reporting without permission under the guise of "fair use." The outcome affects not just copyright law but the economic viability of professional journalism itself. If AI companies can freely train systems on news articles without payment, the financial incentive to invest in original reporting—expensive investigative work, international bureaus, beat reporting—diminishes further. News organizations already struggling with revenue losses face the prospect of subsidizing AI companies' profits through unpaid use of their content. The court's decision on whether to sanction OpenAI for hiding evidence will signal whether companies can obscure their practices while claiming legal protections, or whether transparency and accountability are requirements of the legal process.