“An Icon of Distrust”: Publishing Industry Turns Its Back on Generative AI as Labor and Copyright Disputes Mount

Publishing Industry Reels from Generative AI Controversies and Disputes
Major Technology Companies Ensnared in Litigation over AI Training and Use
Commercial Value and Credibility of AI-Generated Content Continue to Erode

Noise surrounding generative artificial intelligence (AI) is intensifying across the publishing and literary sectors. Disputes have proliferated, with novel contracts being withdrawn over suspected AI use and the results of writing competitions and literary prize deliberations being overturned. Market observers attribute the turmoil to changing perceptions of AI-generated content. AI, once regarded as a symbol of innovation and efficiency, has become a stigma that fuels distrust over authenticity, quality, and accountability.

Novel Publication Abandoned over Suspected AI Use

According to reports published on Aug. 4 by The Wall Street Journal, The Guardian, and other international media outlets, the contract for Nigerian American author Jerry Falade’s crime novel “Call Me, I’ll Hide the Body” was recently withdrawn without warning. The book had been one of the most closely watched debuts in the U.S. and U.K. publishing markets. Minotaur Books, an imprint of Macmillan, secured the publishing rights after offering an advance of more than $2 million in a competitive auction involving 14 publishers and had planned to release the novel in 2028. Film and television production companies had also expressed interest in securing adaptation rights before publication.

The situation reversed after allegations emerged that generative AI had been used in the work. On July 31, the agency that brokered the contract emailed publishers to announce it was withdrawing the manuscript, saying it could no longer verify the process through which the book had been created, from its initial conception to the completed draft. The agency said it had initially accepted the author’s claim that no AI had been used, but lost confidence after his account changed during subsequent discussions. Falade denied the allegations, saying, “The story, characters, and prose are all mine,” and adding that he retained timestamped drafts, notes, and conversation records. He also argued that existing AI-detection programs have high false-positive rates and are unreliable, while none of the editors or publishers involved in reviewing the manuscript had raised concerns about signs of AI use.

Generative AI Upends the Publishing Industry

Similar incidents have become increasingly common across the publishing industry. American author Mia Ballard’s horror novel “Shy Girl,” for example, gained popularity online after being self-published in February last year. French publishing group Hachette recognized the book’s commercial potential, acquired the publishing rights, and formally released it in the U.K. last November. Following publication, however, readers began alleging that its sentences and phrasing showed characteristics associated with AI-generated writing. After conducting an internal investigation, Hachette canceled the U.S. release scheduled for March this year and halted sales of the U.K. edition. Ballard maintains that she did not use AI to write the novel and says an editor hired to revise the self-published edition used AI without her consent.

Japanese publisher AlphaPolis also canceled the print publication and manga adaptation of the winner of its 18th Fantasy Novel Awards in January. The work, “The Unassuming Skill of Tidying Up Is the Strongest,” had won both the grand prize and the readers’ award, but was later found to have been written substantially with generative AI. In New Zealand last November, Stephanie Johnson’s short-story collection “Obligate Carnivore” and Elizabeth Smither’s novella collection “Angel Train” were excluded from consideration for the Ockham New Zealand Book Awards because AI-generated images had been used on their covers. Although both authors said they had either played no role in producing the covers or were unaware of the AI use, the organizers uniformly enforced rules prohibiting AI-generated content in both the text and cover artwork.

AI-Driven Turmoil Engulfs the Literary World

The controversy is spreading across the broader literary world. A recent dispute in South Korea involving the classic literature application “Book Thief” offers a prominent example. Launched late last year, Book Thief attracted users by offering access to hundreds of classic literary works for less than $14, with a lifetime pass priced at approximately $13.90 and an annual pass at about $6.90. Users subsequently began questioning whether AI had been used for the translations, prompting a flood of online complaints stating, “I would not have subscribed had I known the translations were produced with AI.” Book Thief ultimately acknowledged last month that it had used AI in the translation process and apologized to users. The company issued full refunds to existing customers and suspended new registrations and payments.

Lawsuits filed by artists against AI companies over copyright infringement are also proliferating. In 2024, authors including Andrea Bartz, who wrote the bestselling thriller “We Were Never Here,” filed a class-action lawsuit alleging that Anthropic had used their works without authorization to train its AI model Claude. A U.S. court ruled in Anthropic’s favor on the use of lawfully purchased books for AI training, finding that the practice constituted fair use. It nevertheless found that Anthropic’s downloading of books from pirate websites—sites used to share illegal content—could amount to copyright infringement and allowed damages proceedings to continue on that issue. Last month, Judge Araceli Martínez-Olguín of the U.S. District Court for the Northern District of California granted final approval to a settlement submitted by both sides. Under the agreement, Anthropic must place $1.5 billion into a settlement fund for authors and destroy the copyrighted works and copies it downloaded from pirate websites.

Table 1. Generative AI Controversies in Publishing and Literature

Meta Sued over Alleged Use of AI in Layoffs

Disputes over AI use are also intensifying across other industries. Last month, 26 Meta employees filed a lawsuit alleging that the company used AI benchmarks to select workers for dismissal. The plaintiffs claimed that Meta referred to productivity-tracking tools and AI token-usage metrics while cutting approximately 8,000 jobs earlier this year, placing employees with records of absences or leave at a disadvantage. They alleged that workers with disabilities and employees who had taken medical or family-care leave were subjected to discriminatory treatment.

The complaint cited several AI-assisted systems used internally by Meta, including MetaMate, a large language model-based assistant; Second Brain, a tool trained by employees; and productivity scores generated through analysis of keystrokes, screen content, emails, and browser histories. Meta stated in court filings and public comments that all decisions related to this year’s layoffs were made by people and that AI usage was not used to select employees for dismissal or assess performance. A U.S. federal district court recently dismissed the plaintiffs’ claims, finding that there was insufficient documentary or testimonial evidence to establish that AI had intervened in the actual decision-making process or that the process had been improper.

Aversion to AI-Generated Content Deepens

The proliferation of AI-related disputes reflects a rapid shift in public perceptions of AI-generated content. AI use was once treated as evidence of innovation and efficiency. The flood of AI-generated material has since eroded scarcity while deepening concerns over quality and reliability. One market expert said, “In publishing, the fact that a book was written directly by a person is emerging as a new premium, while in art and design, whether a work was produced by a human has a direct impact on its economic value.” The expert added, “The market has begun judging works according to whether a person devoted time and effort to creating them, as well as who produced them and what experiences and intentions informed the work.”

Similar trends are emerging in public administration, finance, and healthcare. During the early development of the AI market, the involvement of AI in these fields was cited as evidence of fairness and objectivity. The absence of discretionary human judgment strengthened confidence in the resulting decisions. Emphasizing AI involvement is now widely perceived as an attempt to evade accountability. Responsibility remains unclear when an AI-generated decision contains errors or causes tangible harm. Improving public perceptions of AI use therefore requires human personnel who provide final approval for AI-generated conclusions and assume responsibility for them.