TechCrunch is reporting that the Seattle Times and Newsday have filed suit against OpenAI and Microsoft, becoming the latest in a growing line of news organizations to take legal action over the alleged use of their published journalism to train artificial intelligence systems.
To understand why this matters, it helps to trace how the news industry arrived here. For most of the past decade, publishers watched their advertising revenue migrate toward platforms that aggregated and redistributed their work without compensation. The legal and contractual frameworks governing that relationship were never fully resolved, but the industry largely adapted by chasing traffic, building social audiences, and striking distribution deals. Generative AI has reopened that wound in a more fundamental way. The argument being made by these plaintiffs, and by others who have sued before them, is not simply that their content was shared without payment — it is that their content was consumed at industrial scale to build a commercial product that now competes directly with the journalism it was trained on. A person who might once have clicked through to a Seattle Times article to get an answer can now receive something resembling that answer from a chatbot, with no visit, no subscription prompt, and no revenue flowing back to the newsroom that did the reporting.
The legal theory resting underneath these suits generally involves copyright infringement. News organizations argue that the scraping of their archives without license constitutes unauthorized reproduction of protected creative work. OpenAI and Microsoft have broadly contested this framing, suggesting that training on publicly available text constitutes a form of fair use — a transformative application rather than direct reproduction. Courts have not yet produced settled precedent on this question, which is part of what makes the accumulation of suits significant. Each new filing adds to the pressure for a ruling, or for a negotiated industry settlement, that could define the relationship between AI developers and content creators for years.
The New York Times filed what became the most prominent of these suits late in 2023, and the list of plaintiffs has lengthened steadily since. That the Seattle Times and Newsday are now joining reflects something worth noting about the composition of the coalition. These are not purely digital-native outlets or niche publications. They are established regional and metropolitan newspapers with decades of archived journalism — exactly the kind of dense, factual, locally-specific reporting that is valuable as training data precisely because it is hard to generate synthetically. Regional papers have also been among the most financially vulnerable parts of the press ecosystem, which raises the stakes on both sides. For publishers already operating on narrow margins, the prospect of AI systems effectively monetizing their archives while simultaneously reducing reader engagement is an existential threat, not an abstract one.
For OpenAI and Microsoft, the consequences of losing these cases, or of facing an unfavorable ruling on fair use, could be significant and structural. It would potentially require renegotiating or licensing vast amounts of historical training data, which would be expensive and logistically complicated in ways that might disadvantage smaller AI developers more than the largest players. Ironically, that dynamic could entrench incumbents like OpenAI and Microsoft even further, since they would be better positioned to absorb licensing costs than newer entrants. Some publishers have already moved toward licensing deals rather than litigation — the AP and several others have reached agreements with AI companies — and a courtroom loss for the plaintiffs could push more of the industry toward that negotiated path.
The likely reading of the current moment is that these suits are functioning as much as leverage as they are as genuine bets on courtroom victory. Publishers want licensing revenue, and litigation is a way of demonstrating that the alternative to a deal is expensive and disruptive. OpenAI and Microsoft, for their part, have strong incentives to settle selectively and avoid a ruling that could set broad precedent. The pattern in technology law has often been that cases of this magnitude either settle before judgment or take many years to resolve, during which the underlying industry continues to evolve rapidly.
What to watch for next is whether any of these cases move meaningfully toward trial or produce significant pretrial rulings on the fair use question. A decision that goes against the AI developers on training data could trigger rapid licensing negotiations across the industry. Equally important is whether more regional publishers join the litigation or whether some begin to peel off toward licensing arrangements, which would signal that the industry's unified front is fracturing. And quietly behind all of this, regulators in the European Union have been developing their own frameworks around AI and copyright that may resolve some of these questions through legislation rather than litigation, potentially reshaping the rules for everyone.
