British Authors Outraged Over Shadow Library Used by Meta for AI Training

British authors expressed feelings of “absolute sickness” upon realizing their book titles have surfaced in a “shadow library” supposedly utilized by Meta for developing artificial intelligence software. One best-selling novelist remarked, “It’s my entire life. The notion that someone in Silicon Valley is appropriating that work to create generic AI replicas… it’s incredibly distressing.”

The search tool for the LibGen database was launched by The Atlantic last week, following the release of court documents related to a lawsuit filed by U.S. comedian Sarah Silverman and other authors against Meta. The company, which owns platforms like Facebook, Instagram, and WhatsApp, is valued at over £1trn. They are accused of infringing copyright laws by using LibGen, a notorious “shadow library” that purportedly houses millions of pirated books, journal articles, and various materials, in the training of its AI software. Meta has denied these allegations, asserting that the suit should be dismissed, claiming it has made “fair use” of the material and its AI system, Llama 3, does not “replicate” the authors’ works.

Previous court documents revealed that internal communications indicated Meta’s CEO Mark Zuckerberg “approved” the use of the LibGen dataset, despite concerns raised by some employees.

The Society of Authors (SoA) condemned Meta’s alleged actions as “appalling,” stating the company “must compensate the rights holders of all the works it has exploited.” Novelist Rowan Coleman, whose extensive portfolio includes the Sunday Times bestseller The Memory Book, shared her distress, saying, “It’s every single book I’ve ever written. I felt absolutely sick. I have no way of knowing how much revenue that has cost me. Like most writers, I struggle with bills; I have three jobs and a mortgage. Yet, tech billionaires are profiting from my work and that of countless others. How can this be justified?”

Coleman emphasized that Meta appears to have opted for a method to acquire “what they needed cheaply and quickly.” Beyond financial implications, she pointed out a more significant concern: “There’s a genuine risk to our profession’s survival. We may reach a point where no books exist for people to pirate — at least none authored by humans.”

Coleman referenced the recent Netflix drama Adolescence, noting that it wouldn’t exist without the dedication of writers laboring for hours. While well-known authors like J.K. Rowling and Stephen King might earn millions, a 2022 survey revealed that the median income for authors in the UK is around £7,000.

Hannah Doyle, a romcom novelist preparing to release her fifth novel, The Spa Break, shared that two of her previous works are included in the LibGen search. She also holds other jobs to supplement her income as an author, stating, “It’s like David and Goliath — how do we assert our rights against these trillion-pound tech giants?” She continued, “This isn’t right; it’s ultimately theft. They are allegedly stealing our work to enhance their AI systems. What future awaits our careers as a result?”

Doyle suggested the outcome could differ if authors had been approached for remuneration. “AI has immense potential benefits in certain fields like medical research, but we urgently need to establish boundaries before it completely takes over.”

Award-winning writer Damian Barr, whose works also appear in the database, posted on Instagram, stating, “Readers and viewers — much of TV, film, and theatre starts with a book — are being subjected to subpar content generated by machines. Creatively, culturally, and financially, AI is robbing us all.”

TV presenter and author Richard Osman, celebrated for his Thursday Murder Club series, commented on X, “Copyright law is straightforward. To use an author’s work, permission is required. Using it without consent is illegal. While it will be difficult for us and other affected industries to challenge Meta, we will certainly try!”

In his article, Atlantic writer Alex Reisner, who developed the LibGen search tool, mentioned that it is “impossible” to determine precisely which portions of LibGen Meta has accessed, given that the database continually expands. His snapshot was created in January 2025, well over a year after the lawsuit claimed the tech giant accessed it, meaning some titles currently visible might not have been available at that time.

The Society of Authors (SoA) is calling on writers in the UK to reach out to Meta and their local MPs. Chief Executive Anna Ganley emphasized, “Rather than seek permission and compensate for copyright-protected materials, AI companies are deliberately choosing to appropriate them in their bid for market dominance.” She described this behavior by large tech firms as shocking, especially as governments currently fail to strengthen and uphold existing copyright protections.

A spokesperson from Meta responded to Sky News, asserting that the company “has developed transformational GenAI open source LLMs that are driving remarkable innovation, productivity, and creativity for both individuals and organizations.” The statement expressed disagreement with the claims made by the plaintiffs, arguing that the full record presents a different narrative. Meta remains committed to defending itself and supporting the development of GenAI for the benefit of all.

In the U.S., authors—including comedian Sarah Silverman, Richard Kadrey, and Ta-Nehisi Coates—have initiated a class-action lawsuit against Meta in California in 2023. They allege that the tech giant unlawfully downloaded digital copies of their books and utilized them without consent or compensation to train AI. This controversy surrounding LibGen reflects a broader discussion on AI and copyright law. The Authors Guild in the U.S. states that legal action is also being pursued against other AI companies for reportedly using pirated books in addition to Meta. Authors have been informed that if their works were used by Meta, they are automatically part of the Kadrey vs. Meta class action lawsuit “without needing to take any immediate action.”

Additionally, in 2023, the Authors Guild alongside 17 authors filed a class-action lawsuit against OpenAI in New York for alleged copyright violations, with notable plaintiffs including John Grisham, George R. R. Martin, and Jodi Picoult. This issue was also a significant factor in the Hollywood strikes of 2023. However, there are divergent views within the creative sector. For instance, Harper Collins agreed to allow “limited use of select non-fiction backlist titles” for training AI models in collaboration with an unnamed tech company. Furthermore, award-winning crime novelist Ajay Chowdhury expressed his acceptance of the technology in a Sky News interview.

A consultation regarding AI copyright law in the UK concluded in February 2023. The proposed changes include creating an exemption to copyright for AI training, which would not require tech firms to obtain licenses for copyrighted material, placing the onus on creators to opt-out if they wish to protect their works. A government spokesperson indicated that the current copyright framework is limiting the creative industries, but changes will only be made once there is assurance of a practical plan that effectively increases control for rights holders, allowing lawful access to materials for training leading AI models in the UK, and enhancing transparency regarding how materials are used.

Nevertheless, many authors and creatives remain unconvinced. Coleman expressed concern that this proposal “opens the door to significant exploitation of individual rights, data, and creative work,” urging the government to reconsider and protect what is a vital aspect of British cultural identity.