USA Today Co., the company formerly known as Gannett, and 19 of its publications sued OpenAI in a New York federal court on 8 October 2026, alleging copyright infringement and seeking more than $250 million in damages. These are allegations in a complaint; OpenAI had not publicly responded in the coverage I read, and no court has ruled. For content owners the useful question is not who wins but what the case says about licensing, AI crawlers and the value of your archive. Facts are as of 9 October 2026, and this is not legal advice.
Key Takeaways
- The suit alleges willful copyright infringement and removal of copyright management information, and seeks over $250 million, per Forbes and local-news reports.
- Everything described below is the plaintiffs' allegation; I found no OpenAI response and no court finding.
- Plaintiffs cite OpenAI's paid agreements with other news organisations as evidence that content use requires a licence; I did not find OpenAI's side of that argument.
- If you publish content, now is the time to decide your AI crawler policy, document your archive and understand what you would license, and at what price.
- Do not treat a lawsuit as a ranking or visibility signal; it is a business and legal story with indirect effects on how AI products cite publishers.
What was filed
According to Forbes, the plaintiffs are USA Today Co., Inc. and 13 affiliated entities, with 19 named publications, and the case was filed on Thursday 8 October 2026 in a federal court in New York. Local public radio coverage from LPM describes the Louisville Courier Journal as a named plaintiff and says the suit was filed in New York federal court. Coverage differs on how many corporate entities are plaintiffs, so I use "USA Today Co. and its publications" rather than a precise count.
What the plaintiffs allege
As reported by Forbes and LPM:
- OpenAI copied hundreds of thousands of articles without permission to train its models and to produce summaries and reproductions in ChatGPT.
- OpenAI knew or should have known this was infringing, which supports a "willful" claim.
- Programs removed copyright management information (CMI) such as bylines and notices.
- Evidence includes publisher content in WebText and in a 2019 Common Crawl snapshot (C4). The reports cite more than 122 million tokens in C4, with 23 million attributed to usatoday.com and 3.1 million to the Courier Journal's site.
- The complaint quotes OpenAI's head of ChatGPT, Nick Turley, calling the situation an "existential threat" to publishers, as reported by Forbes and LPM.
What they want
The reported relief is damages above $250 million (statutory damages of up to $150,000 per willfully infringed work and up to $25,000 per instance of CMI removal), a jury trial, destruction of models and training sets that incorporate the content, and a permanent bar on using the articles. Forbes lists those figures; I have not read the complaint itself.
What is not established
- No court has found OpenAI liable.
- I did not find a statement from OpenAI in the reports I read; LPM and Forbes say it had not immediately responded.
- One aggregator, ExplainX, openly says it reviewed press accounts rather than the complaint, which is a good reminder to read the primary filing where possible.
Where this fits among other cases
This is not the first publisher suit. The New York Times case against OpenAI and Microsoft, filed in December 2023, survived a motion to dismiss in part: Judge Sidney Stein declined to throw it out in March 2025, narrowing but not removing the main copyright claims, according to Nieman Lab's summary of NPR's report. LPM also notes suits by Alden Global Capital newspapers; Axios reported the original April 2024 filing (a search result I did not open in full). Another summary of this week's suit is at NewsBytes, which I have not cross-checked line by line. ExplainX lists further plaintiffs including The Intercept, Ziff Davis and Encyclopaedia Britannica, but I would verify any individual case before citing it.
On the other side of the market sits licensing. A June 2026 overview by the GEO vendor LLM Pulse tallies roughly twenty publisher deals since 2023, including the Associated Press, Axel Springer, the Financial Times, News Corp, Condé Nast and the Washington Post. Most financial terms in that article are journalists' reports rather than confirmed figures, and the source is a vendor blog, so treat it as a map and not an audit. TechCrunch's report on the Financial Times deal is a primary-style example: a non-exclusive licence covering training and display with attributed summaries and links, with undisclosed terms.
Why a marketer should care
I am not a lawyer, and I would not guess the outcome. But the case touches four things that marketing and content teams control.
1. Whether AI products cite you
If publishers and AI companies end up in licensing arrangements, products may favour licensed sources for display. If suits discourage use of unlicensed content, citation patterns could shift. Both are speculation, but they are the kind of shift that matters for anyone tracking AI visibility. I look at the mechanics in what an AI citation is and the effect of source changes in Reddit and ChatGPT citations.
2. Your own crawler policy
Every site owner is making an implicit decision about AI crawlers today. My earlier piece on Google's AI search opt-out controls explains what the main levers do and do not do. Blocking access can reduce your exposure to training, but it can also reduce your visibility in AI answers. There is no free option, so decide deliberately.
3. The value of your archive
The statutory damages logic in the complaint, as described by ExplainX, is per infringed work, so article counts add up quickly. That is a reminder that a well-organised, registered archive is an asset. If you are a publisher, work with counsel on copyright registration and metadata hygiene.
4. Platform dependency
Google's own publisher experiments, for example the AI contribution pilot, and the Penske v Google dismissal show that courts and platforms are still working out where the lines sit. Do not build a business that depends on one ruling.
Comparison: suing versus licensing versus opting out
| Route | What it means | Trade-off |
|---|---|---|
| Sue | Seek damages and an order against use | Slow, expensive, uncertain outcome |
| License | Negotiate payment and attribution terms | Revenue and visibility, but terms may be one-sided or confidential |
| Opt out via robots/controls | Restrict crawlers that respect them | Cheap, but may reduce AI visibility; not all bots comply |
| Do nothing | Content is crawled under current terms | No cost, no leverage |
ExplainX suggests publishers who sue often end up negotiating licences; that is a commentary on pattern, not a prediction for this case.
What to do this week
- Audit your robots.txt and CDN bot rules. List which AI user agents you allow, block or rate-limit, and note the date of your decision.
- Document ownership. Know which pages are original, which are licensed in, and which are user-generated. This matters whether you license or litigate.
- Track AI citations to your site. Log which prompts cite you now, so you can see whether any policy change moves them.
- Decide your price and your red lines for any AI licensing inbound. Even small publishers may receive offers.
- Talk to counsel before making public claims about infringement. Allegations stated carelessly can create their own risk.
- Keep reporting factual. If you write about the case, attribute every claim to the filing or to named outlets, as I have tried to do here.
What would change my view
I would update this post if OpenAI files a response, if a court rules on a motion to dismiss, or if the complaint text shows details the press accounts missed. I also have not verified several figures attributed to the complaint (such as the token counts) against the filing itself.
FAQ
Who sued OpenAI on 8 October 2026?
USA Today Co., formerly Gannett, and publications including USA Today, the Courier Journal, the Detroit Free Press and the Arizona Republic, per Forbes and LPM. Forbes lists 19 publications.
How much are they seeking?
More than $250 million in damages, according to Forbes and local-news reports. That figure includes statutory damages claimed per work and per instance of copyright information removal.
Has OpenAI responded?
In the reports I read, OpenAI had not immediately responded to requests for comment. Check OpenAI's newsroom and court dockets for updates.
Is Gannett still the right name?
LPM describes USA Today Co. as formerly Gannett. I use both names so readers searching either can find this.
Does the lawsuit mean OpenAI will stop citing news sites?
There is no evidence of that. Treat any prediction as speculation. Watch how ChatGPT sources and citations change over the coming months.
Does OpenAI pay publishers?
Yes, it has signed licensing deals with many publishers according to multiple reports, including a non-exclusive deal with the Financial Times covering training and display. Most terms are undisclosed.
Should my site block AI crawlers because of this?
Not automatically. Blocking can reduce training exposure but may also reduce AI visibility. Decide based on your business model, and review the decision each quarter.
Is this legal advice?
No. I am a marketing practitioner, not a lawyer. For rights, registration or licensing questions, speak to an intellectual property lawyer.
Will this affect my Google rankings?
I found no connection between this lawsuit and Google ranking systems. It concerns OpenAI's training and outputs.
Need help planning for AI visibility and content rights?
If you want help deciding how your content should show up in search and AI answers, see my work at younusfardeen.in and reach me through the contact form. I bring 4+ years of marketing experience in SEO, AEO and content strategy, and I will keep the advice practical and sourced.