A group of major publishers and authors has filed a class action lawsuit against Google, claiming the company trained its Gemini AI model on copyrighted works while deliberately concealing the source of that training data. The lawsuit, filed in the U.S. District Court for the Southern District of New York, alleges that Google not only used protected materials without permission but also intentionally removed or altered copyright information to obscure what it had done.

What Exactly Is Google Accused of Doing?

The plaintiffs in the case include major publishing houses Hachette, Cengage, and Elsevier, as well as author Scott Turow and S.C.R.I.B.E., an authors' advocacy organization. According to the lawsuit, Google trained Gemini on copies of books from two specific programs: Google Books, where publishers voluntarily provided works to make them searchable through limited snippets, and Google Play, where books were uploaded for distribution.

The core allegation is that Google repurposed these materials for AI training without ever receiving authorization to do so. The lawsuit states that "Google illegally copied works from all these scope-limited programs for AI training, knowing it lacked authorization to do so." What makes this case particularly notable is the claim that Google deliberately obscured its actions by removing or changing copyright information on these works.

The publishers argue that Google had a long-standing, transparent relationship with them through Google Books and Google Play, but the company violated the terms of that relationship by using the materials for a purpose that was never discussed or agreed upon. The lawsuit also cites an internal Google document that allegedly warned that using copyrighted books for AI training could be "highly problematic for Google" and might result in potential fines ranging from $10 billion to $100 billion.

How Does This Fit Into the Broader AI Copyright Battle?

This lawsuit is one of many copyright challenges facing AI companies. Google, Meta, OpenAI, and Anthropic have all faced legal action from publishers, authors, and other copyright holders over their use of protected works in AI training. The legal landscape remains unsettled, with courts reaching different conclusions about whether using copyrighted material to train AI models qualifies as "fair use" under U.S. copyright law.

Two early court decisions in California have sided with AI companies, ruling that using copyrighted works for AI training falls within the bounds of fair use. However, these rulings do not necessarily predict how other courts will decide similar cases. The Google lawsuit, being heard in New York rather than California, gives a different judge the opportunity to weigh in on these questions.

One significant precedent came when Anthropic was fined $1.5 billion for pirating works used in its AI training, marking the largest payout in U.S. copyright law history. Around 500,000 writers were eligible for payments of at least $3,000 each. However, many authors chose to opt out of that settlement so they could pursue their own legal action against AI companies over AI training practices.

Steps to Understand the Copyright and Fair Use Questions at Stake

• Fair Use Doctrine: U.S. copyright law includes a "fair use" exception that allows limited use of copyrighted material without permission for purposes like criticism, commentary, and education. AI companies argue that training models on copyrighted text falls within this exception, but publishers dispute this interpretation.
• Scope-Limited Licensing: Publishers argue that when they provided works to Google Books and Google Play, they granted permission only for specific, limited uses such as making books searchable or distributing them for purchase. Using those same works for AI training was never part of the agreement.
• Intentional Concealment: The allegation that Google removed or altered copyright information suggests the company may have known its actions were legally questionable and took steps to hide them, which could undermine any fair use defense.
• Precedent Uncertainty: Because earlier California court decisions favored AI companies on fair use grounds, but this case is being heard in New York, the outcome remains genuinely uncertain and could set a different precedent.

The conflict between AI developers and copyright holders remains too nuanced for any single court ruling to settle the matter definitively. Different judges may interpret fair use differently, and the fact that this case involves allegations of deliberate concealment may distinguish it from earlier cases that were decided in favor of AI companies.

What makes the Google case particularly significant is the relationship history between the company and publishers. Unlike situations where AI companies scraped content from the open internet, Google had an established, contractual relationship with these publishers through Google Books and Google Play. The publishers' argument is that Google exploited that trust by using materials provided for one purpose to train AI models for an entirely different purpose, without permission or disclosure.

As this lawsuit proceeds, it will likely influence how courts and policymakers think about the obligations of AI companies when they use copyrighted material for training. The outcome could shape whether AI developers need to seek explicit permission before using published works, or whether they can continue to rely on fair use arguments to justify their practices.