Four major publishers and author Scott Turow sued Google in a New York federal court on Friday, alleging the company copied millions of copyrighted books without permission to train its Gemini AI models. The lawsuit claims Google sidestepped licensing systems and ignored internal warnings that the practice could expose it to billions in fines-a direct challenge to how tech giants source training data for generative AI.
The allegations in the lawsuit
The 60-page complaint argues that Google first obtained books through its Google Books project for limited purposes, then used them as raw material for AI training. It also alleges the company scraped pirated sources and content behind paywalls. "Google willfully sidestepped this longstanding system designed to protect copyrights and compensate authors and publishers through a series of deliberate choices to develop Gemini," the filing states.
Internal documents cited in the suit reportedly warned that using books this way was "highly problematic for Google" and could lead to fines as high as $100 billion. Publishers say they were never informed their works were being used as AI training data. Kirk Sigmon, a technology and IP attorney at KellDann Law, told Al Jazeera that if the books were acquired unlawfully, any fair use defense "would arguably be mooted." Proving what exactly went into a training corpus, however, remains difficult.
The lawsuit follows an earlier attempt by Hachette and Cengage to join a class action brought by authors in 2023. Hachette said in a statement that the scope of the complaint shows "authors and publishers are united in the goal of protecting their valuable intellectual property rights" across fiction, nonfiction, educational works, and scholarly articles.
A wave of copyright lawsuits against AI
This case is part of a broader push by writers and publishers against AI companies. A lawsuit against OpenAI by the Authors Guild and George R.R. Martin is still pending after a judge refused to dismiss it in October. But a separate case against Meta ended differently: a federal judge ruled in 2025 that training AI on copyrighted books qualified as fair use.
Michael Goodyear, an associate professor at New York Law School, said most of these lawsuits follow the same pattern. "The basic claims are you took copyrighted works and used them for training. These were unlawful copies, that's the training argument." Some also challenge the outputs as infringing, he added.
For writers tracking these legal battles, understanding the nuances of AI for Writers has become increasingly relevant, as the outcomes could reshape how creative professionals protect their work in the age of generative AI.
Proving infringement is a challenge
Even when licensing talks happen, the economics often fall short. Oli Huggins, CEO of ExpertEdge and VP of Partnerships at Packt Publishing, said offers currently circulating value a perpetual AI-training license at roughly $10 per title. "The economics remain deeply unattractive," he said.
Huggins also pointed to a fundamental evidentiary problem. "Once the egg is baked into the cake, it is extremely difficult to identify it, quantify its contribution or prove precisely which copy of a book was used," he told Al Jazeera. A model might show familiarity with a work without reproducing enough verbatim text to build a legal case.
Goodyear noted that courts have yet to settle who is liable when a user prompts a model to produce infringing content. "If the user is actively trying to get the model to infringe, that could mean the user is ultimately on the hook rather than the AI system," he said.
Why this matters for writers
The Google lawsuit crystallizes a core tension for working writers: your published work can be ingested, analyzed, and monetized by AI systems without your consent or compensation-and proving it happened is technically daunting. While some publishers are pursuing licensing deals, the per-title fees currently on offer suggest the market is not valuing individual works in a way that sustains a writing career. The outcome of this case will influence whether AI companies must negotiate with rights holders or can continue treating publicly available text as free training material.
Your membership also unlocks:
