Photo composition using Dell laptops and robot hand; photos of laptops Thursday Review

Photo-composition by Thursday Review

Publishing Groups
Join Legal Action
Against Open AI


| published November 2, 2026 |



By R. Alan Clanton, Thursday Review editor





The number of plaintiffs in a major lawsuit has grown larger. The Association of American Publishers (AAP) has joined other professional, publishing, and copyright groups in a legal action asking that a judge in the U.S. Southern District of New York dismiss claims by Microsoft and OpenAI that they engaged in “fair use” when the two tech giants (along with others) used copyrighted materials to train their artificial intelligence mechanisms.

In its amicus brief, AAP has joined the International Association of Scientific, Technical & Medical Publishers, the Independent Book Publishers Association (IBPA), and the Copyright Clearance Center (CCC) in support of the book publishers, authors’ groups, and newspapers already engaged in a legal action over the use of AI—a case which began with an original lawsuit by the Authors Guild.

The single case has grown by way of effectively consolidating with other similar cases and through dozens of other organizations and groups joining with the original case, now commonly known as In RE: Open AI Inc. Copyright Infringement Litigation. There have been numerous lawsuits working their way through the courts regarding the companies which have developed AI and publishers, authors, newspaper writers, publishing associations, and writers.

Though legally complex and controversial in all elements, the core arguments center on the use of the original works of authors and publishers to “train” AI tools, such as Open AI, ChatGPT, Google’s Gemini, Anthropic’s Claude, and other LLMs (Large Language Models), in effect directing the AI tools to create or write books in the style, voice, or narrative tones of original authors. In theory indistinguishable from the authors’ genuine work, these AI-generated books would undermine and dilute the market.

Of equal importance in the legal case, the “training” was done using real authors’ actual work, without the permission of those same authors. In neither instance, according to the plaintiffs, does this constitute “fair use” as previously and widely understood, nor does the use of AI to train itself on creating content fall into any sort of special treatment or exception. Microsoft and other major AI developers have stood their ground by claiming that the synthesis of already-published materials into AI is well within the realm of fair use.

The lawsuit has generated much attention, and the consolidation of numerous similar lawsuits into one mostly merged lawsuit has progressed steadily throughout the summer and fall.

For example: though the company admits no wrongdoing, Anthropic downloaded millions of books and thousands of newspaper articles, stored them as digital copies, and used the material to train Claude in an effort to ramp up Claude’s effectiveness.

That legal contest rolled through the courts until earlier this year, when a federal judge in San Francisco effectively split the difference, ruling that Anthropic did not violate “fair use” when it directed its Claude tool to digest and cull roughly 7 million books and hundreds of thousands of newspaper articles, but the same judge ruled that Anthropic was liable for significant copyright violations since most of that content had been downloaded illegally—pirated using tools such as LibGen and PiLiMi—then stored in massive caches in Anthropic’s digital library.

Hoping to avert a financial disaster if more authors and publishers joined the case, Anthropic agreed to a settlement of $1.5 billion, still stipulating that it had done nothing wrong. After much back-and-forth, that case was finally settled so that eligible authors would receive a payments for any of their own works found to have been pirated for Claude consumption.

But this, too, has proved controversial, since once the math was completed each of those authors would receive between $2,800 and $3,000 per book. In total, there were about 500,000 books covered by that settlement. In addition, authors whose books were pirated by Anthropic but who did not file a claim by the March deadline will receive nothing, nor will the authors of any books that do not possess either an ASIN or an ISBN—meaning almost all of the privately published or self-published material consumed by Claude.

Authors impacted by the case can go to several pages found in the website of the Authors Guild, which has explanations of how the settlement will play out, including very recent news that the payments to authors will go out the week of Thanksgiving. The Authors Guild also explains that some authors may be able to join in one of the other ongoing lawsuits against other AI companies since the Anthropic Settlement contains no clause or contingency regarding other AI tools or LLMs.

Artificial Intelligence and specially the LLMs require massive libraries of reading and consumption—the more, the better. Anthropic, like some of the other tech giants making large steps into artificial intelligence, have sought to minimize their developmental costs by circumventing the costly and potentially slow process of “buying” books for the LLM’s to consume, or seeking explicit permission from publishers and authors. Instead, Anthropic (and allegedly other AI giants) have obtained millions of books through piracy or through other shadowy channels.

This has rankled thousands of authors and writers, and upended the publishing business model, and the Authors Guild makes its position clear.

“Artificial Intelligence machines capable of generating text-based works pose a serious threat to the writing profession and to our literary culture,” an article on the AG website states right up front. “Generative technologies built illegally on vast amounts of copyrighted works without licenses—that is, without giving authors any compensation or control over the use of their works—are used to cheaply and easily produce works that compete with and displace human-authored books, journalism, and other works.”

The Authors Guild has lobbied strongly for guardrails that would avert the possibility that demand for human-generated content drops sharply, and machine-generated content quickly fills that role. Various associations and groups representing journalists are concerned for the same reason. Indeed, already struggling to maintain relevance in a world in which news is now largely created, filtered and curated through social media, reporters worry that AI-generated news will take human intervention out of the equation completely, replacing it with news and information written by machine intelligence. Several U.S. newspapers have already begun pilot programs with AI-generated news items, including the 160-year-old Cleveland Plain Dealer, one of the first major dail newspapers to introduce AI reporting.

The door also swings the other way, and in the last 12 months along there have been at least six notable instances of authors whose latest books have been “outed” for having been allegedly written entirely (or almost entirely) using the tools of artificial intelligence. This has created several high-profile news events in which contracts were canceled and entire books deals scuttled. In August, Minotaur Publishing and McMillan Publishers scrapped a record-breaking $2 million book deal with author Jerry Falade when it was discovered that his new novel, Call Me, I’ll Hide The Body, was likely the product entirely of artificial intelligence. Falade has denied the accusations.

In September, French officials removed the book It Was That Or Die from the prestigious Goncourt Prize shortlist of works by French-speaking-writing authors when his dazzling new work was deemed the result almost entirely of AI.

As recently as mid-October, prize-winning writer Simbarashe Kundizeza—whose short stories won the Caine Prize for African Writing last month—has been accused by multiple sources of using AI to write The Blue Fighter, his autobiographical story of family struggle. On October 6, Kundizeza received his prize, along with a gift of $10,000. Less than a week later an editor and writer at The Atlantic broke the news that Kundizeza’s story was almost certainly 100% the work of AI.

Like Falade, Kundizeza denies any wrongdoing, and has suggested that the AI-detection tools being used are themselves faulty, if not biased. A few authors whose works have been accused of being the work of AI have lashed out at the AI-detection tools being used, suggesting that such tools can be wildly inaccurate and misleading, and often based on inappropriate biases and miscues based on lack of understanding of other languages and other cultures.

Much of the work of detecting AI-generative material is done by a program called Pangram, which is now itself in the spotlight, and the brunt of complaints by some who regard Pangram as flawed. However, most editors and publishers have determined that Pangram’s accuracy is extremely high, and its ability to ferret out machine-writing is in fact strong.

In the meantime, publishers and authors are facing pressures from both directions as AI advances with breathtaking speed and power, and in the view of some within the publishing industry there will be a need for both significant rethinking—including new honest evaluations between authors and their agents and editors, and an industry-wide reboot of how authors plan to remain relevant in a world in which an AI tool can write a novel effortlessly and quickly.



Related Thursday Review articles:

Suspected AI Use in Novel Scraps $2 Million Book Deal; By R. Alan Clanton, Thursday Review staff; August 22, 2026.

AI Safety Has Moved Beyond the Chatbot Era; By Gleb Tsipursky, Thursday Review contributor; September 14, 2026.