xAI sues man accused of using Grok to create child abuse material
Russell Bloodworth faces over 200 criminal charges as multiple civil suits target xAI
Elon Musk's AI company xAI has filed its own lawsuit against a Bentonville photographer accused of using its Grok chatbot to generate child sexual abuse material, marking the first civil case to name the photographer, Russell Bloodworth, as a defendant since his arrest in June.
Bloodworth, aged 43, operated a photography business in Bentonville for almost 15 years before his June 10 arrest. He is now charged with over 200 counts, among which there are possessing and distributing sexually explicit materials featuring a minor and producing deepfake visual media illegally.
The affidavit filed by the investigating officer reveals that Bloodworth employed Grok software to transform legal photographs of his underage clients into sexually explicit AI-generated pictures and videos.
The man entered a plea of not guilty and was released on a $350,000 bond, prohibited from communicating with children outside his family until September 21.
The lawsuit was filed on July 30 in the Northern District of Texas in which xAI alleged that Bloodworth violated the terms of service of Grok by making four separate accounts in March, putting up ordinary pictures of actual persons, including minors, and using “adversarial” prompts to circumvent Grok’s safeguards.
xAI claims that it found out about this itself, filed a CyberTipLine complaint to the National Center for Missing and Exploited Children on April 23, and as a result of this, Bloodworth was identified and arrested.
In light of the indemnification clause in its terms of service, xAI seeks a court order to compel Bloodworth to compensate its legal fees for any suits related to this and that he cannot make a Grok account anymore.
xAI's complaint repeatedly describes Grok as "a powerful, neutral generative artificial intelligence tool", arguing that image outputs are entirely the result of user prompts rather than product design.
The company also points to its enforcement record, saying it suspended 77,723 accounts and filed 100,235 CyberTipline reports in 2026, leading to at least 299 arrests.
