What happened

On 1 September 2026, the US Department of Justice filed a 20-page statement of interest in The New York Times’ copyright lawsuit against OpenAI and Microsoft, urging the judge to rule for the two companies.

The government’s position is that training a large language model on articles isn’t infringement in itself. The model learns statistical relationships between words and ends up doing something different from what the original reporting does, which makes the use highly transformative and therefore fair use. Barring AI training on copyrighted material, it added, could “severely hamper” technological progress. A statement of interest isn’t a ruling, and the judge is free to ignore it. Still, US press coverage described this as the first time the federal government has formally stepped into the current wave of AI copyright cases.

Later that week, on 4 September, The Seattle Times and Newsday sued OpenAI and Microsoft in federal court in Manhattan. Alongside copyright infringement they claim trademark dilution, alleging that the models attribute fabricated content to the two papers.

Europe is running on a separate track. The Commission’s enforcement powers over providers of general-purpose AI (GPAI) models took effect on 2 August. From that date, providers have to keep documentation, put a copyright policy in place and publish a “sufficiently detailed” summary of the content used to train their models, and the AI Office can request information, run evaluations and, where needed, impose fines. In early September, the AI Office sent its first requests for information to more than 30 model providers, covering safety and security on one side and copyright and transparency on the other. The Commission hasn’t said who received them.

Why it matters to you

In the space of one week, the US government spoke up for the AI companies and the EU started asking model makers for paperwork. Both deal with the same end of the pipe: whose content the models took, and how.

Neither Washington nor Brussels touched how AI describes you in its answers. The only place it came up was The Seattle Times’ complaint: a newspaper with its own legal team found AI putting invented content under its name, and the route open to it was a federal trademark-dilution suit. The New York Times filed its own case in December 2023. Nearly three years on, there’s still no judgment.

So if your working assumption has been “AI took our material, and sooner or later the law will deal with it,” the signals from that first week of September aren’t encouraging: Washington has picked a side, and it’s the other one.

Lawsuits run on a clock measured in years. AI answers go out every day. Someone asking ChatGPT, Gemini or Perplexity today who to hire in your field, or whether your product is any good, gets an answer today, built from whatever the model can find right now.

What to do about it

Act, but aim the effort where it can actually land. On the copyright front, a Taiwanese brand has very little leverage, since most companies won’t fund a cross-border lawsuit. Leave that fight to the lawyers, trade associations and big publishers.

What’s left in your hands is how AI writes about you in its answers. AI describes a brand from statements it can find that agree with each other: what your own site says, what coverage and reviews elsewhere say, whether the facts line up across all of them.

When it gets you wrong, nobody tells you. The Seattle Times at least found out for itself. What most brands run into is the other kind of error: AI treating old information, or someone else’s mistaken account, as you. Customers who read it rarely come back to mention it, and the AI sends no notice. The only way to learn which lines about you are wrong is to go looking, and the hard part is tracing them back: AI rarely tells you where it picked something up, the same wrong line resurfaces under a different phrasing or in a different engine, and fixing one source doesn’t mean it’s gone. The later you catch it, the more customers have already read the wrong line.

Whether AI has any line about you wrong right now is something you only find out by asking. Book a demo and we’ll run a round of your industry’s questions live; you’ll spot the wrong lines yourself at a glance.

Further reading