The work of Orna Ross, founder of the Alliance of Independent Authors, and thousands of other independent authors and ALLi members have been identified as part of the vast collection of books used—without permission or payment—to train Meta’s AI models.
“This is a defining moment… Whatever our opinion about, or use of AI in our work, creators can unite around the need to protect copyright law.” — ALLi Founder Orna Ross
A recent investigation by The Atlantic magazine has confirmed what many in the author community have long suspected: the works of thousands of writers—including members of the Alliance of Independent Authors—were used without permission or compensation to train Meta’s artificial intelligence models.
This includes books scraped from the shadow library, Library Genesis (LibGen), which Meta then fed into its LLaMA AI system to develop large language models. Among the authors affected is ALLi Director and Founder Orna Ross, whose fiction, poetry, and nonfiction titles appeared in the dataset.
Ross, a long-standing advocate for author rights and ethical creative entrepreneurship, expressed gratitude to The Atlantic for their ongoing work in uncovering the covert actions of tech giants in relation to copyright infringement.
“This is a defining moment. The author and creative community owes much to The Atlantic for giving us the evidence of what everyone knows has been going on. The foundations of generative AI are being built on unlicensed intellectual property, in the largest unlicensed data grab in history.”
ALLi recognizes that opinions on the use of generative AI across the author and creator community vary widely and the organization has developed an ethical policy around AI for indie authors based around four pillars: clarity, consent, compensation, and curiosity.
“It’s one thing for an author to freely choose to make their work available as open source for discovery or sharing motives,” Ross said. “It’s quite another to have one of the world’s largest corporations deliberately and secretly ride roughshod over copyright law and appropriate creators' property for their own profit–all the while knowing that books generated by AI would compete with those by human authors in the marketplace.”
How Authors Can Check If Their Work Was Used to Train Meta AI
The Atlantic has created a searchable tool that allows copyright holders to look up their name or book titles to see if their work appears in the dataset derived from the Library Genesis (LibGen) shadow library.
To check your titles:
- Visit The Atlantic‘s database tool
- Use the search box to enter your author name or the title of your book
- If your work appears in the results, it was likely part of the dataset
Why This Matters
This isn't just a copyright issue—it’s about creative ethics, author livelihoods, and the future of publishing and other creative industries. Meta is one of several tech giants racing to train AI models on vast swathes of content, much of it scraped without transparency or compensation. And while AI presents opportunities for authors, it must be developed on fair and lawful terms.
ALLi is actively tracking this issue with its members and has renewed its call for clarity, consent and compensation from AI developers, and for licensing mechanisms that respect copyright, intellectual property, and the livelihoods of authors and other creators.
“Whatever our opinion about, or use of AI in our work, creators can unite around the need to protect copyright law,” says Ross. “We can be AI curious and AI positive without giving tech giants carte blanche. We cannot allow the future of creativity to be built on a foundation of furtive, dishonest appropriation.”
For more information about the Alliance of Independent Authors and its position on AI and copyright, visit Selfpublishingadvice.org/ai
ALLi is currently collecting reports from members who discover their books have been used. Authors can contact the Alliance with concerns at info@allianceindependentauthors.org.
You don’t need a subscription to access the database, Cindy, just click on the link in the introduction to the article.
Is there a way to check if my work has been used without subscribing to The Atlantic?