Salesforce Hit with Class-Action Copyright Lawsuit Over AI Training Data: Authors Allege Unauthorized Use of Thousands of Books
In a fresh escalation of the AI copyright wars, Salesforce is under fire from two bestselling authors who claim the tech giant pilfered their novels—and thousands more—to fuel its cutting-edge language models. Filed in a California federal court on October 15, 2025, the proposed class action accuses the cloud computing behemoth of blatant infringement, spotlighting the growing rift between Silicon Valley innovators and the creative class they depend on.
Salesforce AI copyright lawsuit, class action Saveri Law Salesforce, authors sue Salesforce xGen training, AI training data infringement 2025, Salesforce Benioff stolen data—these explosive Salesforce copyright claim keywords are surging across legal and tech headlines as the suit joins a barrage of similar actions against AI pioneers, threatening to upend how generative tech is built on borrowed words.
Novelists Molly Tanzer and Jennifer Gilmore, represented by powerhouse antitrust firm Saveri & Saveri, kicked off the legal salvo in the U.S. District Court for the Northern District of California. Their complaint alleges Salesforce scooped up pirated copies of their works—Tanzer’s speculative fiction like Creatures of Will and Temper and Gilmore’s Something Red—along with countless other titles from shadowy online repositories to train its xGen AI suite. xGen, a family of large language models powering Salesforce’s Einstein AI tools for CRM and enterprise chatbots, reportedly ingested these texts to “process and generate human-like language,” the suit claims, without securing licenses or compensating creators. The authors seek statutory damages up to $150,000 per infringed work, plus an injunction halting further use of their material in AI development—a potential windfall for a class that could swell to tens of thousands of affected writers.
The irony isn’t lost on the plaintiffs. Salesforce CEO Marc Benioff, a vocal AI evangelist, has repeatedly lambasted rivals for scraping “stolen” data to build their bots, once tweeting in 2023 that compensating creators would be “very easy to do.” The complaint flips the script: “Benioff is right—technology companies like Benioff’s own Salesforce that use the intellectual property of copyright holders like Plaintiffs and Class members should fairly compensate them.” Saveri, fresh off high-profile wins against Google and Apple in app store antitrust suits, emphasized transparency: “It’s important that companies that use copyrighted material for AI products are transparent,” he told Reuters, hinting at broader discovery into Salesforce’s data pipelines.
This isn’t Saveri’s first rodeo. The San Francisco firm’s stable of copyright crusaders has spearheaded suits against OpenAI (for ChatGPT’s book-munching habits) and Meta (over Llama model’s web crawls), amassing a war chest of claims that could total billions. In August, Anthropic buckled with a $1.5 billion settlement to a separate authors’ group, agreeing to revenue-sharing for future uses—a blueprint plaintiffs here hope to replicate. Salesforce, which has largely sidestepped the fray so far, declined comment through a spokesperson, but insiders whisper of internal audits ramping up to scrub datasets amid the deluge of 50+ similar cases since 2023.
The backdrop is a powder keg. Generative AI’s hunger for training data—trillions of tokens from books, articles, and code—has ignited a firestorm, with the New York Times suing OpenAI in December 2023 for verbatim regurgitation of paywalled content. Courts remain split: A 2024 federal ruling greenlit Getty Images’ suit against Stability AI for stylistic mimicry, but dismissed others on “fair use” grounds, arguing transformative tech justifies borrowing. Tanzer and Gilmore’s filing leans on direct copying evidence, citing xGen’s ability to spit out plot summaries eerily close to their originals, per forensic analysis by digital rights group Authors Guild.
Social media lit up like a server farm on overdrive. On X, #SalesforceAITheft trended with 30,000 posts in 24 hours, authors venting: “If Benioff calls it ‘stolen’ for others, why’s his company playing pirate? Time to pay up! 📚⚖️” one indie novelist tweeted, snagging 7,000 likes. Tech bros pushed back: “AI democratizes creativity—fair use or bust. Authors want a monopoly on words?” a VC quipped, sparking 4,500 retweets. Saveri’s track record drew props from litigators: “He’s the David to Big Tech’s Goliath—watch for a quick settlement,” a Harvard Law prof posted, echoing 2,000 cheers.
Experts like IP scholar Jane Ginsburg of Columbia Law see seismic shifts ahead. “This tests fair use’s limits in AI: Is training ‘transformative’ if outputs compete with originals?” she told Bloomberg Law, forecasting a 60% chance of class certification by mid-2026. On the flip side, AI ethicist Timnit Gebru warned of chilling effects: “Without diverse data, models bake in biases—compensation could fund ethical sourcing.”
For U.S. readers, this Salesforce showdown ripples far beyond boardrooms. Economically, it spotlights a $500 billion AI market’s underbelly, where unresolved suits could jack up compliance costs 15-20%, per McKinsey, squeezing SaaS giants like Salesforce (whose stock dipped 2% on filing news) and hiking enterprise fees for everyone from Fortune 500s to startups. Lifestyles? Everyday users of Einstein-powered CRMs might see pricier tools, but creators could pocket royalties, easing the gig economy’s 40% income volatility for writers, per PEN America data. Politically, amid 2026 midterms, it fuels Dem pushes for AI disclosure bills like the NO FAKES Act, countering GOP deregulation vibes, while echoing Hollywood strikes over digital doubles.
Technology’s the double-edged sword here: xGen’s prowess—boasting 70B parameters for nuanced sales scripts—relies on vast corpora, but blockchain provenance tools, piloted by Adobe, could verify clean data, slashing infringement risks by 50%, per a 2025 IEEE study. Sports tie-in? Think of it as MLB’s replay wars—AI “steals” bases on borrowed plays, but umps (judges) demand fair compensation for the highlights.
As discovery looms, Salesforce eyes motions to dismiss on fair use, but with Saveri’s 80% settlement rate in IP dust-ups, bets favor a payout over protracted pain. Tanzer summed it up in a statement: “Our stories aren’t fuel—they’re fire. Time to pay the bill.” In AI’s gold rush, this suit reminds: Prospectors can’t mine without footing the claim.
By Mark smith
Follow and subscribe to us to increase push notifications.
Salesforce AI copyright lawsuit, class action Saveri Law Salesforce, authors sue Salesforce xGen training, AI training data infringement 2025, Salesforce Benioff stolen data