Elon Musk’s AI company xAI filed a federal lawsuit in Texas on July 14, 2026, targeting a South Carolina man named Terry Harwood for allegedly using the Grok chatbot to produce sexually explicit deepfakes involving minors. The case is being watched closely across the tech and AI industry, because it represents one of the first times an AI company has gone on legal offense against a user for generating child sexual abuse material through its platform.

What xAI says Harwood did

According to the lawsuit, Harwood uploaded non-sexual images of both adults and minors into Grok and manipulated the AI into generating explicitly sexual deepfakes from those source images. The allegations describe a deliberate effort to circumvent safety guardrails that xAI says were built into the product specifically to prevent this kind of abuse.

Harwood is not a stranger to law enforcement. He was arrested in February 2026 on separate charges of sexually exploiting minors, well before the xAI lawsuit was filed. The civil case against him now runs parallel to whatever criminal proceedings stem from that earlier arrest.

xAI is seeking unspecified monetary damages and a permanent ban preventing Harwood from accessing any xAI product. No specific dollar figure has been attached to the civil claim.

Why this lawsuit matters beyond one defendant

The Harwood case does not exist in isolation. xAI disclosed that it suspended 52,222 accounts in 2026 and filed 73,604 reports with the National Center for Missing and Exploited Children. Those reports led to at least 244 arrests.

The AI safety and regulatory backdrop

The deepfake CSAM problem predates large language models, but generative AI has made it substantially worse. Earlier tools required meaningful technical skill. Modern multimodal AI systems, including Grok, can transform ordinary photographs into explicit content with minimal friction, which is exactly what the Harwood lawsuit describes.

In the US, the DEFIANCE Act, signed into law in 2024, created a civil right of action for victims of nonconsensual intimate deepfakes. Several states have passed their own legislation. But most of these frameworks focus on the people who create and distribute the material, not on the AI platforms that may have facilitated production.