An Irish writer at a desk, surrounded by books, looking thoughtfully at a computer screen displaying AI text.

Irish Writers Secure AI Settlement

Irish writers are set to receive a share of a €1.31 billion settlement from a copyright infringement case against the AI company Anthropic. This agreement marks a significant development for creative professionals in Ireland.

An Irish writer at a desk, surrounded by books, looking thoughtfully at a computer screen displaying AI text.

The Irish Writers Union confirmed that this settlement, the first of its kind against tech companies for using copyrighted material to train large language models, will distribute millions of euros among Irish authors. Conor Kostick, the Union's disputes officer, emphasized the personal importance of these funds for many writers, including himself. He also hopes this case will pave the way for further legal action against international tech companies that he believes have "cavalierly [stolen] copyrighted works" to build their generative AI models.

The class action lawsuit, filed in 2024 by Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, involved nearly half a million works by the time it concluded. The complaint primarily focused on Anthropic's use of the LibGen online library, a platform known for providing access to millions of titles outside traditional copyright regulations. The compensation specifically addresses Anthropic's use of LibGen, not the broader issue of unlicensed copyrighted works for training. This distinction matters because a U.S. judge suggested the latter might fall under "fair use," a defense unlikely to succeed under European law.

The $1.5 billion (€1.31 billion) settlement amounts to roughly $3,000 (€2,630) for each book confirmed to have been copied. While this flat rate may not drastically change the financial situation for bestselling Irish authors like Sally Rooney, Marian Keyes, or Colm Tóibín, it represents a meaningful sum for many others. Kostick himself could receive up to €7,500 for his three books, though the final amount will be split between authors and their publishers.

A notable condition for compensation was that a book had to be registered for copyright in the United States on the settlement date. This requirement will unfortunately exclude many Irish books, as U.S. copyright registration is not standard practice for all authors there. Despite this, Kostick believes the case sets a crucial precedent. He noted that it has since become clear Anthropic was not the only AI developer training its models on pirate databases. The real challenge, he added, lies in the power imbalance between individual writers and the vast resources of big tech companies.

Ivan O'Brien of O'Brien Press expects his company to receive over €10,000 for their covered books. He mentioned that while hundreds of their books were likely used in Anthropic's data build, only about a dozen met the specific criteria for this class action. For O'Brien Press, it's a "nice rather than business amending" sum, but the acknowledgment of the principle that "stealing stuff is wrong" is a welcome victory.

Children's author Oisín McGann will have four of his books covered, potentially earning him up to €10,000. He expressed satisfaction at the prospect of earning from works he hadn't expected further income from. However, he also highlighted his frustration that 12 other books were excluded due to the U.S. copyright registration requirement, which he noted isn't typically necessary for Irish authors.

Barry Scannell, a partner at William Fry specializing in copyright and AI, offered a cautious perspective. While acknowledging the settlement is likely the largest ever for copyrighted material, he suggested it might inadvertently strengthen the idea in U.S. law that using non-pirated books for AI training constitutes "fair use." This interpretation, he warned, could lead tech companies to aggressively acquire and scan physical books, believing they don't need to pay royalties.

The copyright situation in the European Union is, according to Scannell, "very different" but also a "disaster" for both rights holders and those seeking to use copyrighted works. Samantha Holman, chief executive of the Irish Copyright Licensing Agency, sees the case as a positive development, drawing attention to the existence of "shadow libraries." She hopes it will encourage other companies to negotiate, though she remains realistic about the challenges of dealing with tech giants like Meta and Google, known for prolonged legal battles.

This settlement for Irish writers is more than just a financial payout; it's a significant step toward defining intellectual property boundaries in the age of artificial intelligence. It highlights the ongoing tension between technological advancement and creators' rights. As AI continues to evolve, the conversations and legal precedents set by cases like this will be crucial in shaping a fair future for artists and writers worldwide. Understanding international copyright law is becoming ever more important, and resources like the World Intellectual Property Organization (WIPO) offer valuable insights into these complex issues. This is a story we'll be following closely, as the implications for Ireland's vibrant literary community are profound.

Share this content: