Another week, another pair of newspapers dragging OpenAI and Microsoft into court — and I’m starting to wonder if anyone in the AI industry even flinches anymore when these lawsuits drop.
What Actually Happened
On September 4, 2026, The Seattle Times and Newsday filed a federal copyright and trademark complaint against OpenAI and Microsoft in the U.S. District Court for the Southern District of New York. The core claim is familiar territory by now: these publications allege that OpenAI and Microsoft used their copyrighted content without permission to train AI models, and that this unauthorized scraping directly harms their journalism businesses.
If you’ve been following the AI space for more than five minutes, you know this is not the first time we’ve seen this play out. Major newspapers have been lining up to take their shots at OpenAI. And yet, here we are. The models keep getting trained. The products keep shipping. The lawsuits keep piling up in a neat little stack on some federal judge’s desk.
My Honest Take I use them daily. I depend on them. And I’m telling you plainly: the copyright question isn’t going away, and the AI industry’s strategy of just absorbing legal punches until courts sort it out is cynical at best.
When I test a new AI model, I’m looking at output quality, accuracy, reasoning ability — all the stuff that matters to end users. But every single one of those capabilities was built on a foundation of training data. And a meaningful chunk of that training data came from newsrooms like The Seattle Times and Newsday, staffed by reporters who actually leave their desks and cover city council meetings, local crime, school board fights, and community issues that nobody else bothers with.
Strip-mining that work to build products worth billions? That’s a business model question that deserves a straight answer, not a legal rope-a-dope.
The Justice Department Threw a Curveball
Here’s where things get particularly uncomfortable for the pro-publisher crowd: the Justice Department has sided with the tech companies on this, arguing against the copyright infringement framing. That’s a significant signal. When the federal government weighs in on the side of OpenAI and Microsoft, it shifts the political gravity of the entire debate.
Does that mean the newspapers are wrong? Not necessarily. It means the legal terrain is genuinely contested and the outcome is far from predetermined. Courts could still rule in favor of publishers. But having the DOJ in your opponent’s corner is not exactly a confidence booster if you’re The Seattle Times’ legal team.
Why This Matters for AI Users
If you use AI agents and tools — which, if you’re reading agnthq.com, you probably do — this fight matters to you more than you think. The quality of AI outputs is directly tied to the quality of training inputs. If newsrooms continue to shrink because their content gets vacuumed up for free, the information ecosystem that feeds these models degrades. You end up with AI trained on AI-generated content trained on thinner and thinner source material. That’s a quality death spiral nobody in the industry wants to talk about publicly.
I’ve tested models that confidently spit out local news information with fabricated details. I’ve seen outputs that clearly echo the reporting style and factual structure of specific publications without attribution. The connection between training data provenance and output reliability is real, and as a reviewer, I can’t ignore it just because the tools are useful.
What Should Actually Happen
I’m not a lawyer, and I’m not going to pretend that copyright law neatly answers every question about AI training. But I think we need licensing frameworks that actually work — not the performative deals that a few major outlets have signed while hundreds of smaller newsrooms get nothing. The Seattle Times isn’t The New York Times. Newsday isn’t The Wall Street Journal. These are regional publications with finite resources, and they’re fighting for survival while their content feeds systems they never consented to participate in.
OpenAI and Microsoft can afford to pay for what they use. They choose not to when they can get away with it. That’s not a legal argument — that’s just an observation about how power works.
Where This Leaves Us
As someone who spends every day evaluating AI products, I want these tools to be as good as possible. I also want them to be built on something more defensible than “we took everything and dared people to sue us.” The Seattle Times and Newsday lawsuit is case number whatever in a growing list, and each one chips away at the idea that this is a settled question.
For now, I’ll keep reviewing the tools honestly, including the uncomfortable parts. That’s the job.
🕒 Published: