Background
Authors E. Molly Tanzer and Jennifer Gilmore filed a proposed class action on 15 October 2025 against Salesforce, Inc., asserting that its xGen large-language-model initiative used their copyrighted books without authorisation. They seek statutory damages, declaratory relief, and an injunction. The pleadings allege direct, vicarious, and contributory infringement under the U.S. Copyright Act. No substantive ruling has yet been issued.
AI interaction
The complaint states that Salesforce “trained..." Large Language Models (‘LLMs’), including xGen, using “the Infringed Works.” It also alleges that the defendant’s training datasets included material obtained from “Bibliotik,” “Z-Library,” and other “pirate libraries.” (ECF No. 1). Taken together, these assertions frame the case around the provenance of datasets and the legality of reproducing literary works for model-training purposes, positioning the dispute as an early U.S. test of copyright liability for enterprise-scale AI training.