Google faces a fresh copyright lawsuit from three major publishers—Hachette Book Group, Cengage Learning, and Elsevier—over allegations that the technology giant trained its artificial intelligence models on copyrighted books without authorisation or compensation.
The lawsuit, filed in a US federal court, represents the latest legal challenge to Big Tech’s data acquisition practices for AI development. According to TechCrunch AI, the publishers claim Google systematically copied and ingested their copyrighted works to build commercial AI products, including its Gemini large language models, without obtaining licences or paying royalties.
The case follows a pattern of mounting legal pressure on technology companies over AI training data. The New York Times sued OpenAI and Microsoft in December 2023, whilst authors including John Grisham and George R.R. Martin filed similar claims against OpenAI earlier that year. This lawsuit, however, marks a significant escalation by targeting Google specifically and involving publishers that collectively control thousands of academic, educational, and trade titles.
Hachette Book Group, owned by French media conglomerate Lagardère, publishes authors including Malcolm Gladwell and James Patterson. Cengage Learning dominates the educational publishing market with textbooks and digital learning platforms. Elsevier, part of RELX Group, controls a substantial portion of academic and scientific publishing, with over 2,500 journals and 39,000 book titles in its catalogue.
The publishers argue that Google’s actions constitute wilful copyright infringement on a massive scale. They contend that whilst Google has entered licensing agreements with some publishers—including News Corp and the Associated Press—the company proceeded to use works from non-consenting publishers without permission. The lawsuit seeks both injunctive relief to prevent further unauthorised use and monetary damages.
Google has not yet filed a formal response to the complaint. The company has previously defended its AI training practices under the doctrine of fair use, arguing that transformative use of copyrighted material for machine learning constitutes legitimate research and development. This defence remains untested in the courts, with no definitive legal precedent yet established in the United States or United Kingdom.
Business Impact
The lawsuit creates immediate uncertainty for Google’s AI development roadmap and could force the company to either negotiate expensive licensing deals or fundamentally alter its training data sources. If the publishers prevail, Google could face statutory damages potentially reaching billions of pounds, given the scale of alleged infringement.
Publishers stand to gain significant leverage in negotiating AI licensing agreements, potentially creating a new revenue stream as print sales continue declining. Academic publishers like Elsevier, which already operate on high-margin subscription models, could extract substantial fees from AI companies requiring access to specialised knowledge domains.
Competitors including OpenAI, Anthropic, and Meta face similar exposure, though this lawsuit specifically targets Google. The case could establish precedent affecting the entire AI industry’s approach to training data acquisition, potentially advantaging smaller AI companies that have proactively secured licensing agreements or those developing models on exclusively licensed or synthetic data.
The legal uncertainty may accelerate investment in alternative training approaches, including synthetic data generation and human-curated datasets with clear provenance. Companies providing licensed training data, such as Shutterstock and Getty Images, could see increased demand.
What Comes Next
The case will likely take years to resolve through the US court system, with potential appeals extending the timeline further. Legal observers will watch closely for Google’s formal response and any motion to dismiss based on fair use defences.
Parallel developments in the European Union, where the AI Act includes provisions for transparency in training data, may influence the case’s trajectory. UK copyright law, whilst similar to US doctrine, lacks an equivalent fair use provision, potentially creating different outcomes for similar cases filed in British courts.
The lawsuit’s outcome will substantially shape whether AI companies can continue accessing vast corpuses of copyrighted material without permission, or whether they must negotiate a complex web of licensing agreements that could slow development and concentrate power amongst well-capitalised incumbents. For publishers, the case represents perhaps their best opportunity to establish that their intellectual property retains value and legal protection in the age of artificial intelligence.
