They sued Meta last month. Now Hachette, Elsevier, and Cengage are taking aim at Google. The complaint dropped on July 10 in federal court argues that the search giant stole millions of books. They didn’t just read them. They devoured them to fuel Gemini.
Why authors and publishers are targeting Google’s Gemini
The core argument is blunt. Google “cashed in” on relationships with content creators only to bypass permission. The plaintiffs claim Google brazenly copied their copyrighted works. This training data allowed the Gemini model to compete with human writers at a speed that feels unnerving.
It raises a tough question: should big tech pay for the foundation of its AI? The lawsuit suggests yes. They allege Google scraped web content and Google Books archives. No ask. No payment. Just raw extraction for commercial gain.
Generative AI copyright claims keep popping up for this reason. AI needs massive datasets to function well. Much of the internet’s high-quality text is protected. When companies harvest it without clearance, legal friction follows. It’s almost guaranteed.
“Copyright law applies to AI companies… with the same force as every other.”
Hachette Book Group is leading this charge alongside Cengage and Elsevier. Author Scott Turow joined as well. You might recognize these names. They are part of a growing coalition. Back in May, this group teamed up with Macmillan and McGraw-Hill. Together, they filed a nearly identical suit against Meta. The strategy is consistent. Hit the models at the source.
The landscape of AI copyright disputes in publishing
This isn’t isolated noise. It reflects a broader rot in how media values human creation. Hachette actually canceled a book recently. Shy Girl by Mia Ballard vanished from shelves after users alleged AI assistance. The backlash was immediate. It violated publisher guidelines and inflamed readers. The message was clear. AI usage isn’t welcome without transparency.
Google has faced heat before. Disney sent a cease-and-desist last December. They were mad about Nano Banana and other video models. Disney claimed those tools rode for free on their intellectual property. Mickey Mouse doesn’t give discounts to AI trainers.
But courts haven’t been uniform in their response. Two major rulings against Anthropic and Meta favored the AI developers. Yet those judges left the door cracked. Future cases could swing differently. Context matters. Scale matters. Permission matters.
Which companies are at risk? Likely the ones relying on unlicensed web scrapes. The Hachette complaint argues Google copied the class’s works specifically. This implies a widespread issue affecting many creators, not just a few large names.
What this means for digital content and AI models
The stakes are high. If courts side with publishers, the cost of training large language models skyrockets. Google hasn’t commented yet. Their legal team usually moves fast on these. But for now, silence prevails.
The publishing world is watching. Every ruling reshapes the industry. Will we see licensing deals emerge from litigation? Probably. Will prices rise? Almost certainly.
For now, the argument stands. Copying is not the same as accessing. Training is not reading. If you want to build a tool that competes with human creativity, you pay for the human creativity that fed it. That is the thesis of this lawsuit. And Google has plenty of resources to fight it.
