Google faces fresh copyright lawsuit over Gemini AI training data

MUMBAI: The next chapter in the AI revolution may be written in a courtroom. Google is facing a fresh copyright battle after a group of publishers and bestselling author Scott Turow accused the technology giant of using millions of copyrighted books without permission to train its Gemini artificial intelligence models. A proposed class action lawsuit has been filed in a federal court in New York by Hachette Book Group, Cengage Learning, Elsevier, Scott Turow and his publishing company S.C.R.I.B.E., alleging that Google unlawfully copied copyrighted books to develop its AI systems. The plaintiffs are seeking class action status, an injunction to halt the alleged conduct and unspecified damages. According to the complaint, Google copied millions of copyrighted works that had originally been provided to services such as Google Books for limited purposes. The lawsuit alleges that the company subsequently repurposed this material to train its Gemini large language models without obtaining permission from authors or publishers. The complaint further argues that Gemini now produces AI-generated content that competes directly with the original works of authors, raising concerns over the commercial impact of generative AI on the publishing industry. The publishers also contend that Gemini is capable of generating books at a speed and scale that human writers cannot match, while mimicking the distinctive creative styles and expressive choices of individual authors. The latest filing adds to an expanding wave of legal challenges confronting AI developers over the use of copyrighted material for model training. In May, the same group of publishers and Turow filed a similar lawsuit against Meta, making comparable allegations regarding the company’s AI training practices. The case also follows a series of significant court rulings that have begun shaping the legal boundaries of AI development in the United States. In September, a US judge approved a $1.5 billion settlement between Anthropic and several authors over allegations that the company copied books to train its Claude AI model. While the court held that AI training could qualify as fair use because of its transformative nature, it ruled separately that other uses involving allegedly pirated material did not receive the same protection. Meta also secured a partial legal victory last year when a federal judge in San Francisco ruled that its use of copyrighted material constituted fair use in a lawsuit brought by comedian Sarah Silverman, author Ta-Nehisi Coates and other writers. The latest lawsuit underscores the growing tension between AI developers and the publishing industry as courts continue to grapple with whether the use of copyrighted material to train generative AI models falls within the limits of existing copyright law. As companies race to build increasingly powerful AI systems, the outcome of these legal battles could help define the future relationship between artificial intelligence and intellectual property.