Complaint alleges xAI built Grok on a dataset containing known child sexual abuse material and released image tools with no meaningful safeguards, enabling the mass production of sexualized images of real people, adults and children alike.
SAN JOSE, CALIFORNIA — Girard Sharp, Liberty Law, and Marsh Law Firm filed a nationwide class action today against xAI in the U.S. District Court for the Northern District of California, alleging xAI trained Grok on data including child sexual abuse material (CSAM) depicting real children and that Grok generated CSAM based on those images.
The lead plaintiff, identified in the complaint as Jane Doe, was a young child when she was repeatedly raped to produce CSAM for predators. Those child sexual abuse images have been included in the CSAM Hash List maintained by the National Center for Missing & Exploited Children (NCMEC) since the early 2000s, when NCMEC first identified her in the material. The complaint alleges that same material was part of the dataset xAI used to build Grok’s image and video generating capabilities; that Grok generated new sexually explicit images of her likeness; and that her abuse material also served as a basis for CSAM Grok generated of other victims.
“Our client has lived for nearly two decades knowing that images of the worst thing that ever happened to her are circulating among predators online, and that they can resurface at any moment,” said Sarah London, a partner at Girard Sharp LLP and counsel for the plaintiff and the proposed class. “xAI must be held responsible for knowingly training its models on images of the horrific abuse she suffered, and on the abuse images of every other survivor in this class.”
xAI released Grok Imagine, an image and video model built to create sexually explicit content from real photographs of real people and marketed its “spicy mode” as a selling point. On December 31, 2025, Elon Musk personally promoted Grok’s ability to remove clothing from real people’s photographs on X, sparking a viral trend in which users directed Grok to digitally undress real women and children. Researchers at the Center for Countering Digital Hate estimated that from December 29 to January 8, Grok generated approximately three million sexualized images, including roughly 23,000 depicting apparent children. When faced with public backlash over the deluge of highly sexualized AI-generated deepfakes, xAI limited Grok’s image and video generating capabilities to paid subscribers, where reporting from WIRED found content was “far more graphic” than what was publicly posted on X.
“This case is about xAI’s decision to profit directly from the mass production of AI-generated CSAM,” said Micha Liberty owner of Liberty Law and counsel for the plaintiff. “Every other major AI company screens abuse material out of its training data and blocks nonconsensual deepfakes from being generated. xAI did neither. And when faced with the flood of AI CSAM, the company moved its image and video generating capabilities behind a paywall, turning abuse into a source of revenue.”
Competing companies offering AI image generating models, including OpenAI and Meta, block explicit content, prevent nonconsensual sexualized deepfakes of real people, and filter abusive material out of their training datasets. The complaint alleges xAI did none of that adequately. Musk pushed back against safety restrictions and cut the company’s internal safety team to roughly two or three people for most of 2025. Days before launch, a senior xAI employee publicly stated the company “urgently” needed safety staff; asked whether xAI does safety work, the employee answered: “working on it.” The complaint further alleges Grok’s system prompt was written to assume good intent whenever a request included words such as “teenage” or “girl.”
“Possessing CSAM is a crime, producing CSAM is a crime, and distributing CSAM is a crime. xAI did all three,” said Margaret E. Mabie, partner at Marsh Law Firm PLLC. There is no artificial intelligence exception to federal child protection laws and xAI must face the consequences of its reckless actions.”
The case brings claims under Masha’s Law 18 U.S.C. § 2255, and a related claim under 18 U.S.C. § 2252A(f), both of which give CSAM victims a civil right of action, on behalf of every person in the United States whose childhood images Grok altered into CSAM, a class that counsel believe numbers at least in the thousands. The lawsuit seeks damages and an order requiring xAI to stop generating, possessing, or transporting this material and to destroy what it has already created.
The case is Jane Doe v. xAI Corp., Case No. 5:26-cv-09016, in the U.S. District Court for the Northern District of California, San Jose Division. The plaintiff and the proposed class are represented by Dena C. Sharp, Adam E. Polk, Sarah R. London, Andrew R. Kaufman, and Muriel Kenfield-Kelleher of Girard Sharp LLP; Micha Star Liberty of Liberty Law; and Margaret E. Mabie of Marsh Law Firm PLLC.
For a copy of the complaint or to speak with counsel, contact Adrianna Harper at adrianna@rebuttalpr.com.