xAI Sues Its Own User Over Grok-Generated Child Abuse Images While Facing a Separate Class Action for the Same Problem

Elon Musk's AI company is now on both sides of the same courtroom problem.

xAI filed a lawsuit against Terry Wayne Harwood, a South Carolina man already facing eight felony charges for allegedly possessing and distributing child sexual abuse material, according to Reuters and The Verge. xAI claims Harwood "knowingly and intentionally used Grok to circumvent safeguards, alter nonconsensual images, and generate and distribute CSAM," violating the company's policies. The company wants Harwood to pay for its legal costs, including expenses from any lawsuits filed by his victims, and wants a judge to permanently ban him from using Grok.

Harwood was arrested in February. But xAI suing one user doesn't erase the bigger legal problem the company is facing: a class-action lawsuit that says Grok's own guardrails failed repeatedly, and that xAI didn't handle the aftermath properly.

The Class Action Keeps Growing

The original lawsuit was filed by three Tennessee teenagers in March, according to CyberScoop. Since then it's expanded twice. Two new plaintiffs, referred to as Jane Doe 4 and Jane Doe 5, joined via an amended complaint filed this week, according to weku (NPR). Stability AI, maker of the Stable Diffusion image generator, was also added as a defendant.

Jane Doe 4's case is the most disturbing detail in either story. According to weku and CyberScoop, her stepfather uploaded a single photo of her taken when she was about 11 years old, lying on a couch. Using Grok, he generated roughly 7,000 sexually explicit images and videos from that one photo, including depictions of her performing sexual acts on men, some with explicit captions. He allegedly picked Grok specifically, the complaint says, because it was "less restrictive than other AI models."

The lawsuit alleges xAI sent exactly one tip to the National Center for Missing and Exploited Children about Jane Doe 4's case, in February, and that tip only included the original authentic photo, not any of the roughly 7,000 AI-generated images. When law enforcement asked for the generated images and the perpetrator's IP address to help identify him, xAI allegedly didn't respond, according to both weku and CyberScoop. Her stepfather was eventually arrested and charged, then died by suicide two days later.

Jane Doe 5's account, per CyberScoop, involves an adult male relative of a classmate who used Grok to turn an eighth-grade graduation photo into abuse material that was then traded online. The man was arrested and charged. The images are reportedly still circulating.

What's Proven, What's Alleged

The Harwood case is a lawsuit xAI itself filed, built on criminal charges already pending against him. It's a company using civil litigation against someone it says abused its product, and it's the first time xAI has done this, according to The Verge.

The class-action allegations against xAI, by contrast, are unproven claims in a complaint, not a verdict. Attorney Annika Martin, representing all five plaintiffs, told NPR the lawsuit is about establishing "public nuisance" liability: "this is not just something that is problematic for our clients... this is something that is a scourge on society." That's her characterization, made on behalf of paying clients in active litigation, not a finding of fact.

Companies are legally required to report suspected CSAM to NCMEC. Whether xAI's single, incomplete tip regarding Jane Doe 4 constituted an adequate report, or a failure, is exactly what the lawsuit will have to prove. NPR notes many electronic service providers submit high volumes of reports to NCMEC that lack "sufficient or actionable" detail, suggesting this may be an industry-wide reporting quality problem, not necessarily one company acting in bad faith.

xAI's press office did not respond to a request for comment, according to CyberScoop.

The Larger Pattern

This isn't xAI's first brush with the deepfake problem. After the company rolled out a "spicy" mode for Grok and added AI image-editing capabilities last year, The Verge reported a wave of sexualized deepfakes followed, some involving minors. That reporting led Apple to reportedly consider pulling Grok from its App Store, according to an earlier Verge report from April. Musk responded publicly at the time that "anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content."

Two unresolved questions now sit at the center of this: whether xAI's Harwood lawsuit signals it will start aggressively pursuing individual bad actors going forward, and whether the class action can prove xAI's NCMEC reporting fell short of what the law requires. Neither has been decided by a court. Both cases proceed on separate tracks, with Stability AI now also a defendant in the class action alongside xAI.