A group of major publishers and authors filed a class action lawsuit against Google in the US District Court for the Southern District of New York on Monday, alleging the company used their copyrighted works to train its Gemini AI platform without permission. The plaintiffs include Hachette, Cengage, Elsevier, novelist Scott Turow, and the writers’ organisation S.C.R.I.B.E. The lawsuit accuses Google not only of using the works without authorisation but of deliberately removing or altering copyright management information – metadata identifying authorship and rights status – in order to conceal that Gemini was trained on what the plaintiffs describe as stolen materials. An internal Google memo cited in the complaint reportedly warned that using publisher-provided copyrighted books for AI training was “highly problematic for Google,” flagging between $10 billion and $100 billion in potential penalties. Gemini has more than 650 million monthly active users. YourNewsClub identifies the copyright management information removal allegation as the most legally significant element of Monday’s complaint: it describes not inadvertent use of copyrighted material during model training – which is the basis of most AI copyright cases and the subject of the fair use debate – but deliberate concealment of that use, which could displace fair use as the central legal question and introduce wilfulness findings that carry multiplied statutory damages.
The lawsuit’s structure exploits a specific prior relationship. Publishers provided Google with access to their books specifically to power Google Books, a programme designed to generate searchable text excerpts without displaying complete content. The plaintiffs now argue that Google trained Gemini on copies of those same books – as well as Google Play uploads – without ever obtaining permission for that substantially broader commercial use. That argument frames the case as a consent violation rather than simply a fair use dispute: publishers agreed to one use, Google allegedly performed a different and more expansive one.
Two California court decisions in 2026 ruled that use of copyrighted works for AI training constitutes fair use. The Southern District of New York is not bound by California precedent, and its judges have historically applied copyright law with attention to market harm to creative industries well-represented in the Manhattan legal community. The New York venue selection by plaintiffs is deliberate. YourNewsClub notes that Anthropic was separately fined $1.5 billion in what TechCrunch describes as the largest payout in US copyright history – though the case against Google differs structurally from the Anthropic settlement, particularly given the Google Books prior relationship that did not exist in Anthropic’s case.
Google has not commented on the lawsuit. The case is styled as a putative class action on behalf of a broader class of authors and publishers whose works may have been used without permission. The same publisher coalition – overlapping substantially with Monday’s plaintiff list – recently filed a nearly identical complaint against Meta in the same Southern District courthouse over Meta’s training of its Llama models. The dual filings in Manhattan suggest the publishing industry has converged on a coordinated litigation strategy using the same jurisdiction, the same counsel, and substantially the same plaintiff class, rather than pursuing individual cases against each AI company in different venues.
Jessica Larn, who studies macro-level technology policy and infrastructure impact of AI, draws the structural copyright question: “Publishers consented to digitisation for discoverability; they did not consent to training. Whether that distinction matters under copyright law is the question no court has yet directly answered at this scale.” Owen Radner, who models digital infrastructure as energy-information transport systems, frames the AI training market implications: “Google’s exposure is complicated by its prior digital relationship with publishers. Having operated Google Books for two decades under a specific consent framework and then allegedly using that same content pipeline for training is a different legal posture from a company that simply crawled the web.”
Your News Club maps the Manhattan venue selection as the most strategically consequential decision in how Monday’s lawsuit was structured, since a New York federal court ruling on AI training fair use would carry weight across every AI company currently under similar claims in courts across the country. The Southern District’s judge will approach the question without being bound by the California decisions that have so far tilted toward AI companies, and any finding on the Google Books consent argument would introduce a new dimension that the existing California precedents do not resolve.
The publishers filing simultaneously against Meta in the same courthouse suggests a coordinated legal strategy designed to force multiple AI companies into discovery simultaneously, potentially producing internal documents from multiple defendants that illuminate industry-wide practices around copyrighted training data. YourNewsClub calls any Google filing in response to the complaint as the first substantive indication of how the company plans to argue the consent-versus-fair-use question that Monday’s lawsuit has explicitly placed at the centre of its theory.