Salesforce Hit with Class-Action Lawsuit from Saveri Law Firm Over Unauthorized Use of Books in AI Training: Transparency & Compensation Demanded
In the latest salvo of the escalating war between tech giants and content creators, Salesforce, the cloud-computing behemoth powering enterprise AI tools, finds itself in the crosshairs of a proposed class-action lawsuit accusing it of pilfering thousands of copyrighted books to fuel its xGen AI models. Filed on October 15, 2025, in the U.S. District Court for the Northern District of California, the complaint—led by powerhouse antitrust firm Saveri & Company—alleges blatant infringement that could reshape how corporations justify “fair use” in the age of generative AI.
The suit, brought by novelists Molly Tanzer (author of speculative fiction like Creatures of Will & Temper) and Jennifer Gilmore (The Sorcerer’s Heir series), claims Salesforce scraped pirated copies of their works and countless others from shadowy online libraries to train xGen, its suite of large language models designed for natural language processing in CRM and customer service applications. “Salesforce’s AI empire was built on the backs of uncompensated authors,” the 45-page filing asserts, estimating the dataset included over 100,000 titles without licenses or permissions, violating the Copyright Act’s core protections. The plaintiffs seek damages up to $150,000 per infringed work, injunctive relief to halt xGen’s distribution, and class certification for all U.S. authors whose books were allegedly ingested—potentially encompassing millions in liability.
Joseph Saveri, the firm’s managing partner and a veteran of landmark cases against Google and Meta over ad tech monopolies, didn’t hold back in a post-filing statement: “It’s important that companies using copyrighted material for AI products are transparent. It’s also only fair that our clients are fairly compensated when this happens.” Saveri & Company, fresh off a $1.5 billion settlement with Anthropic in August 2025 for similar claims, positions this as a “pattern of exploitation” in AI development, where datasets like Books3—flagged by The Atlantic as a “piracy trove”—underpin models without creator consent.
The focus keyword “Salesforce Saveri Law AI copyright lawsuit” encapsulates the urgency of this clash, intersecting with AI training data infringement, class action & author compensation, xGen model controversy, tech copyright battles 2025, and fair use in generative AI that have surged in legal dockets and X debates since the complaint dropped.
The Allegations: From Shadow Libraries to Silicon Valley Servers
At issue is xGen-SF, Salesforce’s open-source family of models (100M to 7B parameters) launched in May 2024, touted for outperforming rivals in tasks like code generation and chatbots but now scrutinized for its “vast, diverse” training corpus. Tanzer and Gilmore’s suit details how Salesforce allegedly accessed their novels—Tanzer’s gothic fantasies and Gilmore’s urban thrillers—via illicit sources like Z-Library, a notorious ebook pirate site shuttered in 2023 but whose archives linger in dark web echoes. The authors argue this “systematic copying” not only deprived them of licensing fees (averaging $5,000-10,000 per title in AI deals) but enabled xGen to “reproduce” stylistic elements, like Tanzer’s intricate world-building, in outputs that mimic protected expressions.
Ironically, Salesforce CEO Marc Benioff has publicly decried the practice, tweeting in June 2025 that compensating creators for AI training data is “very easy to do” and criticizing peers for “stolen” inputs. Yet the suit paints the company as hypocritical, claiming internal memos (redacted in the filing) reveal deliberate use of “gray market” datasets to cut costs amid a $34 billion R&D spend in fiscal 2025. A Salesforce spokesperson, reached by Reuters, declined comment, citing ongoing review.
This fits a broader blitz: Since the New York Times’ 2023 exposé on AI “data poisoning,” over 50 suits have targeted OpenAI (Sarah Silverman et al.), Meta (a class of 10,000+ authors), and Stability AI, with courts split on fair use defenses. A July 2025 federal ruling in Andersen v. Stability AI greenlit claims of “expressive copying,” bolstering Saveri’s odds here.
Legal Heavyweights Weigh In: ‘A Wake-Up Call for Enterprise AI’
Saveri & Company, with its antitrust roots (think the $1.2 billion Epic v. Google win), brings formidable firepower, but experts see this as a bellwether for B2B tech. “Salesforce isn’t a consumer-facing ChatGPT; it’s embedded in Fortune 500 workflows—proving infringement here could cascade to licensing mandates across CRM,” said IP litigator Mark Lemley of Stanford Law in a Bloomberg Law op-ed. On the defense side, Perkins Coie’s AI counsel predicts a “fair use fortress,” arguing transformative outputs like xGen’s analytics don’t “supplant” books but enhance enterprise value—echoing Grok’s 2024 dismissal of similar claims against xAI.
X (formerly Twitter) lit up with polarized takes, #AISteals trending as authors like @MollyTanzer vented: “My words built their billion-dollar bots—time for royalties, not excuses,” netting 15K likes. Tech boosters countered with @BenioffFanClub’s “Innovation needs fuel—fair use is the engine,” sparking 8K replies debating ethics vs. progress. A viral thread from @TechPolicyWatch broke down xGen’s dataset overlaps with Books3, amassing 100K views and calls for SEC probes into undisclosed risks.
Impacts on U.S. Creators, Consumers, & the AI Economy
For American readers—especially the 200,000+ U.S. authors scraping by on $6,000 median incomes—this suit strikes at the heart of creative survival amid AI’s “great extraction.” Economically, a win could unlock $10B+ in retroactive deals, per Authors Guild estimates, stabilizing publishing’s 20% revenue dip since ChatGPT’s debut, while pressuring Salesforce’s $271B market cap (down 2% post-filing). For businesses, it signals audit time: xGen powers tools used by 150K+ companies; tainted models risk contract breaches and data breaches.
Lifestyle ripple? Aspiring writers in Brooklyn lofts or Midwest MFA programs eye eroded markets, where AI “ghostwriters” undercut gigs—pairing with broader job fears as 85% of Fortune 500 firms adopt Salesforce AI. Politically, with Biden’s FTC eyeing “AI monopolies” and Trump’s deregulation whispers, it amps calls for a “copyright czar” in Congress. Technologically, it spotlights “ethical scraping” tools like Adobe’s Firefly, but with xGen’s open-source code, forks could evade injunctions.
User intent skews to strategy: Authors query Saveri’s intake form for class opt-ins; execs scour compliance checklists via Gartner reports; innovators hunt licensed datasets on Hugging Face. Salesforce’s legal team, per insiders, eyes mediation, with a case management conference set for December 10.
As Salesforce Saveri Law AI copyright lawsuit, AI training data infringement, class action & author compensation, xGen model controversy, tech copyright battles 2025, and fair use in generative AI dominate, this filing accelerates the reckoning: Tech’s AI gold rush can’t run roughshod over creators’ rights.
In summary, Saveri & Company’s class action accuses Salesforce of infringing copyrights by training xGen on pirated books, demanding transparency and damages in a suit that could redefine AI ethics. Looking ahead, with discovery looming and fair use motions inevitable, it promises a courtroom clash that might force Big Tech to pay up—or pivot to paid partnerships—in the creativity-vs.-computation tug-of-war.
By Sam Michael
Follow and subscribe to us to increase push notifications.
Salesforce Saveri Law AI copyright lawsuit, AI training data infringement, class action author compensation, xGen model controversy, tech copyright battles 2025, fair use generative AI, Molly Tanzer Jennifer Gilmore suit, Joseph Saveri antitrust AI, Salesforce Benioff AI data criticism, Books3 piracy dataset