Why Did USA Today Co. Sue OpenAI for More Than $250 Million, and What Does It Mean for Coverage Cited in ChatGPT?

Who Sued Whom, and When?

USA Today Co., the company formerly called Gannett, which took its current name on November 18, 2025 (USA Today Co. release), filed suit with 13 companies it owns on October 8, 2026 in the U.S. District Court for the Southern District of New York. The caption lists 14 plaintiffs, including Gannett Satellite Information Network, The Courier-Journal, Inc., Detroit Free Press, Inc. and Phoenix Newspapers, Inc. It names seven defendants: OpenAI Foundation, OpenAI GP, OAI International, OpenAI OpCo, OpenAI Global, OAI Corporation and OpenAI Group PBC.

Paragraph 2 says the plaintiffs together own copyrights in content from 19 publications, among them USA TODAY, The Tennessean, Indy Star, The Arizona Republic, the Detroit Free Press, The Columbus Dispatch and The Des Moines Register. Nineteen is the number of publications; fourteen is the number of plaintiffs. The complaint runs to 79 pages and demands a jury trial. Law Commentary reports the same court, date, defendants and length. Press Gazette's tracker dates its entry October 9 and gives the claim as "$250m" (Press Gazette); the filing itself is dated October 8 and says "in excess of."

What Does the Complaint Allege, and What Does It Ask For?

The complaint has three counts: copyright infringement under 17 U.S.C. § 501, vicarious copyright infringement, and removal of copyright management information under the Digital Millennium Copyright Act. The plaintiffs allege that OpenAI scraped or otherwise acquired their content to train GPT models, that the models "memorized" articles, and that OpenAI's products output summaries and near-verbatim copies. The complaint reproduces prompts the plaintiffs ran on GPT-5.6, which it says retrieved and summarized specific articles, with full versions in an exhibit. It also alleges that OpenAI used text extractors that remove bylines and copyright notices (Complaint ¶¶8, 14, 123, 204). The complaint itself notes that OpenAI has added safeguards against verbatim reproduction, which it calls "litigation-inspired" (¶8).

On money, paragraph 13 says the plaintiffs "seek damages in excess of $250 million" and notes that the law allows up to $150,000 for each willful infringement plus up to $25,000 per violation for removing copyright management information. The prayer for relief asks, without a figure, for statutory and compensatory damages, restitution, disgorgement, a declaration, a permanent injunction and fees. It also asks for destruction under 17 U.S.C. § 503(b) of "all GPT or other LLM models and training sets" that incorporate the plaintiffs' content (Complaint ¶13 and Prayer for Relief). A "$250 million lawsuit" is shorthand for a damages figure stated in the body of the pleading.

What Has a Court Decided, and What Has OpenAI Said?

No court has ruled on the merits. A complaint is one side's account, written to persuade, and OpenAI has not yet filed a response in the material we read. The table keeps allegation and status apart.

We found no OpenAI statement on this case. The Law Commentary and Press Gazette reports we read quote none, and we could not open OpenAI's own website to check for one. A ruling on fair use in another case would not, by itself, decide this one. This is not legal advice.

What Evidence Does the Complaint Offer?

The main evidence is dataset counts and output tests. Paragraph 102 says the plaintiffs' content makes up more than 160,000 entries in WebText, the dataset OpenAI built for GPT-2. Paragraph 106 says their domains account for more than 122 million tokens in "C4," a filtered English-language subset of a 2019 Common Crawl snapshot (Complaint ¶¶102, 106).

The chart shows how concentrated the alleged material is: usatoday.com accounts for 83,266 of the "more than 160,000" entries, and no other domain reaches 13,000.

Dataset counts, if accurate, show that pages were collected. They do not on their own establish that a later model reproduces those pages or that the use was unlawful. Those are the questions the court would have to decide, and OpenAI will have its own account of them.

Why Would a Publisher Sue OpenAI While Licensing to Others?

The plaintiffs say licensing is the point. The complaint argues that OpenAI "knows a license is required" because it has licensing agreements with "over a dozen news and media organizations, including the Associated Press, Axel Springer, The Atlantic, and Vox Media" (¶149). It also says the plaintiffs license content to other media companies "with clear guidelines" and that unlicensed use undermines those arrangements (¶¶184-185). Both are the plaintiffs' arguments.

Press Gazette's tracker, updated October 12, 2026, shows the same group on both sides of the market (Press Gazette).

A publisher can therefore be a licensor to one AI product and a plaintiff against another. We did not read any of the contracts, and their terms are not public. Our wider list of deals is in which publishers have AI licensing deals.

Does a Publisher's AI Stance Change How Its Coverage Reaches AI Tools?

It can, in two ways the complaint describes. First, access controls: the plaintiffs say their sites block OpenAI's crawlers in robots.txt and send detected AI crawlers to a page stating "Crawling and scraping of our site is not permitted." (¶145). Second, terms of service: the complaint quotes USA Today's terms as barring use of site content to develop AI systems, including for "retrieval augmented generation" (¶144).

We did not check live robots.txt files or OpenAI's crawler documentation for this guide, so we cannot say what any publication's settings permit today or how ChatGPT handles them. Our guide to whether news sites block AI crawlers covers how those files work. Our inference, not a platform rule, is that coverage in an outlet that blocks a given AI company may be less reachable by that company's products than coverage in an outlet that licenses to it. The sources we read do not show this.

The complaint also argues that a user who has read an AI answer, "even—or especially—with attribution" to the publication, "has less reason to visit the original source" (¶121). That is a plaintiff's argument, but it is a useful reminder that being cited is not the same as being read.

What This Means for Anyone Buying a Placement

• Treat the suit as information about the outlet. The complaint concerns the plaintiffs' own journalism. It does not address how ChatGPT handles third-party, sponsored or press-release pages on these sites; on that distinction, see whether a press release on USA Today's site is USA Today coverage.
• Ask the outlet three questions before you buy: which AI companies it licenses to, which crawlers it blocks, and whether its terms restrict AI use. Our recommendation, not a rule: get the answers in writing and re-check them during a campaign.
• Do not buy a placement to get cited by ChatGPT. No outlet or vendor can promise an AI citation, a ranking or permanence (can you pay to appear in a ChatGPT answer). A paid article or sponsored post is advertising: it should be labeled and disclosed, and its links are typically nofollow or sponsored (what the FTC requires).
• Expect positions to move. Press Gazette's tracker logged Reach joining Google's news AI pilot on October 12, 2026, three days after it logged this suit. An outlet's stance today may not hold for a whole campaign.

Methodology and Limitations

Four limits apply. First, we read the complaint as filed (Document 1) through a copy hosted by a legal-news blog; we did not see the court docket, so anything filed after October 8 is unchecked. Second, the Reuters report was paywalled and we read only its headline, and we could not open OpenAI's website, so OpenAI's position appears here only as reported by Law Commentary. Third, the licensing table relies on Press Gazette's tracker and one company release, not on contracts. Fourth, we tested no AI product and checked no live robots.txt file. This is not legal advice.

Conclusion

One publisher group is litigating against one AI company while licensing to others. Whether its claims succeed is unknown: the complaint is an advocacy document, OpenAI has not yet answered it in court, and no judge has weighed the evidence. For a buyer, the uncomfortable conclusion is that how coverage travels into AI answers is partly governed by contracts, crawler settings and court outcomes that neither the client nor the outlet fully controls. Buy placements for the readers and credibility they bring now, and treat any AI visibility as a possible side effect that no one can promise.

Frequently Asked Questions

Sources

This guide is built from the complaint in USA Today Co., Inc. v. OpenAI Foundation (Document 1, filed October 8, 2026), Press Gazette's AI deals and lawsuits tracker, Law Commentary's report of the filing and USA Today Co.'s November 2025 name-change release, all read on October 12, 2026. Statements about what the complaint says come from the complaint itself; statements about licensing deals come from Press Gazette. The inference about how a publisher's stance affects AI tools and the questions to ask an outlet are house recommendations, not rules. We did not read the docket after filing, OpenAI's own statements or crawler documentation, live robots.txt files or any contract, and we did not test any AI product.

1. court filing USA Today Co., Inc. et al. v. OpenAI Foundation et al., No. 1:26-cv-08892 (S.D.N.Y.), Complaint, Document 1, hosted copy (https://chatgptiseatingtheworld.com/wp-content/uploads/2026/10/COMPLAINT-USA_Today_Co_Inc_et_al_v_OpenA.pdf). Parties, publications, counts, damages language, prayer for relief, WebText and C4 figures, crawler, terms and licensing allegations. Read October 12, 2026.
2. trade press Press Gazette, Who's suing AI and who's signing latest: Reach latest Google AI news partner, USA Today Co sues OpenAI. Suit entry and date, Perplexity, Meta and Microsoft entries, Reach–Google entry. Read October 12, 2026.
3. trade press Law Commentary, USA Today Hits OpenAI With $250 Million-Plus Copyright Lawsuit Over ChatGPT. Court, date, defendants and length; related-case request; OpenAI's reliance on fair use in other suits and the plaintiffs' response. Read October 12, 2026.
4. official USA Today Co., Gannett Rebrands to USA TODAY Co.. Name change effective November 18, 2025. Read October 12, 2026.