An analysis by the Center for Countering Digital Hate found that, during an 11‑day period between Dec. 2025 and Jan. 2026, Grok created more than 3 million sexualized images — at least 23,000 of which appeared to depict children.
The plaintiffs’ central allegations
A class action lawsuit filed in the U.S. District Court for the Northern District of California alleges that xAI trained Grok’s synthetic deepfake “nudify” capabilities on real images and videos of child sexual abuse material (CSAM). The lead plaintiff, named in the complaint as Jane Doe 1, is described as "an identified victim of child pornography tracked by the Federal Bureau of Investigation’s Child Exploitation Notification Program." The complaint says Doe was a pre‑school aged child when her abuser began creating and distributing CSAM and that the media depicting her has circulated online since at least the early 2000s.
The suit alleges that material containing the plaintiff has "well known hash‑values" and that those hash‑values "have shown up in deepfakes created with Grok and spread on X." It quotes the allegation that “xAI, using Grok, has generated images depicting Plaintiff and the child pornography series in which she is the victim.” The plaintiffs are seeking monetary damages and injunctions to stop further creation of sexualized deepfakes and CSAM; the complaint says the proposed class contains "thousands" of members with similar claims.
Evidence cited: NCMEC, FBI notifications and public analysis
The complaint ties the plaintiff’s ongoing harm to law enforcement tracking and public reporting. It notes that “hundreds of thousands of files” including material of the plaintiff have been included in law enforcement submissions to the National Center for Missing & Exploited Children (NCMEC), and that, as part of the FBI’s Child Exploitation Notification Program, she still receives alerts when images related to her abuse surface online.
Separately, the Center for Countering Digital Hate’s analysis — cited in the complaint — measured the spike in sexualized images output by Grok during the 11‑day window and flagged thousands that appeared to depict children. Those figures form the empirical backbone of the plaintiffs’ claim that Grok has produced and distributed CSAM at scale.
Audit-ready is a season. It shouldn't be.
Evidence in spreadsheets, controls drifting between audits, frameworks multiplying on flat headcount. Nubivance runs continuous compliance on Rapid7 Cyber GRC - SOC 2, HIPAA, ISO 27001, PCI, CMMC.
End the scramblexAI’s defenses, Musk’s public statements, and training‑data concerns
xAI has publicly said it built guardrails to prevent Grok from creating deepfake sexualized images, but the lawsuit contends those protections are "very weak," that they "diverge from standard industry best practice," and that they are easily circumvented. The complaint describes Grok’s filter behavior in technical terms: the system rejects prompts when the chatbot detects “clear intent by the user in the phrasing of their request,” but it argues that “indirect or euphemistic prompts can easily slip past a text‑based filter, so if the model retains the underlying capability to generate sexual or abusive content, some volume of CSAM becomes effectively inevitably generated by the system.”
The complaint also points to Grok’s integration with X and to xAI’s terms of service, which treat content posted on X as training material; plaintiffs say that any CSAM posted to the site over the past year was likely ingested by the model. On Jan. 14 — after the CCDH period cited in the complaint — Elon Musk wrote on X that he was “not aware of any naked underage images of Grok. Literally zero.” Musk also implied that users, not the model, bore primary responsibility for illicit images while stating the model “will refuse to produce anything illegal” if asked. The suit says xAI did not adequately curb the resulting harms. A request for comment from xAI was not returned.
Criminal probes, private suits, and policy fights
The complaint comes amid a wave of criminal investigations, private litigation, and legislative responses tied to Grok’s nudification capabilities. The suit notes another private lawsuit earlier this year alleging a man used Grok to produce thousands of deepfake CSAM images of his stepdaughter and that xAI withheld information from law enforcement that could have helped to identify the perpetrator. According to that complaint, days after authorities traced and seized the material on the wte, he committed suicide.
Policy pushback has followed: in July, Elon Musk sued Minnesota Attorney General Keith Ellison over a new state law that bans nudification technology and imposes a $500,000 fine for each instance where an AI tool is used to generate or alter an image in a sexually explicit way. Musk’s lawsuit argues the Minnesota ban violates the First Amendment, and the state challenge underscores how lawmakers and companies are colliding over how to regulate AI image creation.
What this means for victims, law enforcement, and regulators
- Victims and survivors: Plaintiffs seek monetary damages and injunctions to stop new deepfakes and CSAM; the complaint emphasizes ongoing harm from images that have circulated for decades and continue to surface through FBI and NCMEC tracking.
- Law enforcement and investigators: The complaint relies on FBI notifications and NCMEC submissions to document the provenance and recurrence of illicit media, and prior suits claim that platform cooperation (or its absence) affected criminal investigations.
- Regulators and state officials: The Minnesota law and Musk’s July challenge illustrate competing legal strategies — statutory bans with steep fines versus constitutional defenses — that will shape whether state‑level restrictions on nudification tools persist or face judicial rollback.
The case filed in the Northern District of California will test how civil remedies such as those preserved under Masha’s Law — which the complaint invokes and which "protects civil legal remedies to victims of child pornography and online exploitation" — apply when plaintiffs allege generative AI models absorbed and reproduced their abuse imagery. The complaint frames the dispute as both a claim of past harm and a request for court orders to stop further damage; how the court weighs evidence about model training, guardrails, and platform responsibilities will determine whether a new legal path opens for victims seeking redress against AI creators and platforms.
