Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Hachette Book Group, Cengage Learning, Elsevier, and author Scott Turow are suing Google in US federal court in New York, alleging that the Silicon Valley technology giant violated the law in training its Gemini AI models.
“Google willfully ignored a long-standing system designed to protect copyright and pay authors and publishers by deliberately choosing to create Gemini1,” the nearly 60-page complaint was filed on Friday.
Recommended Articles
list of 4 itemsend of series
The lawsuit alleges that Google copied books as sources through Google Books, saying that the company used the books it “acquired for limited purposes related to Google Books and other Google services”. They also claim that Google “downloaded content from the entire web, including content known to pirates and behind legitimate paywalls”.
It also claims that Google copied the works without permission to train its AI models and continues to do so, even though the works appear to be inconsistent with existing agreements.
The court said the company was well aware of the legal risks, saying internal documents that warned against using textbooks to train AIs were “extremely difficult for Google,” and could lead to $100bn in fines.
“At no time did Google inform authors and publishers that Google was copying their articles as a source for developing and training AI models,” he said.
“The idea, in short, is that any fair use (a legal doctrine that allows the use of prohibited material without permission for reasons such as reporting, education, and research on a limited scale) the argument that Gemini has can be challenged because they obtained the books illegally,” Kirk Sigmon, the first partner who focuses on technology and IP law, told AlJanzen Law. “It’s an interesting story with a lot of complications, not least because it’s hard to prove what was or wasn’t in the training team.”
The lawsuit follows an attempt by Hachette Book Group and Cengage in February to join a lawsuit that already existed with the group of authors in 2023.
“The magnitude of the complaint confirms that authors and publishers are united in protecting their valuable intellectual rights in fiction, non-fiction, children’s books, memoirs, poetry, as well as academic works and scholarly articles that create thousands of educational and research activities,” said Hachette in a statement following the complaint.
Google did not respond to Al Jazeera’s request for comment.
The number of cases
This is far from the first lawsuit brought against an AI company by authors and publishers for alleged copyright infringement.
There is also a pending lawsuit against OpenAI brought by authors including George RR Martin, author of Game of Thrones, and the Authors Guild. In October, a federal judge rejected OpenAI’s attempt to dismiss the lawsuit.
However, a separate lawsuit brought by a group of writers against Meta did not go in the writers’ favor. In 2025, a group of authors led by Richard Kadrey reported that Facebook and its parent Instagram used copyrighted books to train its AI models. A federal judge ruled that AI training meets the law’s “fair use” requirement.
“They (private lawsuits) all follow the same road map,” Michael Goodyear, an associate professor at New York Law School, told Al Jazeera.
“The one that says you took copyrighted books and used them to teach.” These were illegal copies, that’s the teaching point.
But Oli Huggins, CEO of ExpertEdge and VP of Partnerships at Packt Publishing – a publisher whose positions were used in the Anthropic fraud case and who told Al Jazeera that they were approached by AI companies who want to pay to learn more about the company – said that the problem is that when the information is used to teach, it is difficult to verify if there is a copy.
“Verifying what happened inside the sample is another big challenge. Once the egg is baked into the cake, it’s very difficult to identify it, calculate what it gave or confirm exactly which book was used.” An example can show a good knowledge of the work without repeating a clear statement to confirm the method of proof that the claimant wants,” Huggins told Al Jazeera.
Huggins said the licensing deals currently being offered are not sustainable for publishers.
“Finances remain abysmal. Current offerings can have a perpetual license for AI courses for about $10 per chapter.”
There are many lawsuits in content-driven industries, including news and music, against AI companies for copyright infringement as well.
CNN has been sent a case against Perplexityclaiming that the company illegally copied more than 17,000 stories to train its models, which it created “similar or very similar to CNN’s content”, according to the complaint filed in May.
Last week, 17 media organizations, including The New York Timescriticized OpenAI for refusing to testify in a lawsuit brought by the newspaper in 2023, claiming that the AI company led by Sam Altman had violated their rights by teaching ChatGPT.
In the music industry, Hagens Berman, a well-known law firm, sued the AI music maker, claiming that it trained independent musicians without their consent.
In January, Universal Music Group sued Anthropic, claiming that the Dario Amodei-led company infringed their rights by using 20,000 songs to teach its Claude version without permission.
But there are looming questions about who is responsible for what appears to be plagiarized content, according to Goodyear, which the courts have yet to answer.
“If the user is trying to get the model to be broken, it could mean that the user is on the hook instead of the AI,” Goodyear said. “This remains an open question that the courts in the United States have not grappled with.”
