xAI Sues Photographer It Says Bypassed Grok’s Safety Guardrails

xAI says Grok initially rejected prompts seeking to sexualize images of children. Its federal lawsuit alleges that Bentonville photographer Russell Bloodworth repeatedly altered those prompts until he bypassed the chatbot’s safeguards and used the system to create sexually explicit images involving children. Families have separately sued xAI over alleged sexualization of their children’s images.

An alleged route around the controls

xAI filed the case in the U.S. District Court for the Northern District of Texas. It alleges that Bloodworth used Grok to make sexually explicit images involving children and, after the system refused requests under its moderation rules, repeatedly submitted adversarial prompts—altered instructions intended to evade those restrictions.

The complaint frames the alleged conduct as deliberate violations of xAI’s terms of service and acceptable-use policy. Those civil allegations have not been decided by a court.

A CyberTip and contract claims

xAI says it submitted a CyberTip to the National Center for Missing and Exploited Children on April 23. The company says the report was forwarded to the Bentonville Police Department on June 8 and that its reporting led directly to Bloodworth’s arrest.

The lawsuit also seeks a declaration that Bloodworth violated xAI’s policies and enforcement of an indemnification provision, a contract term that can require one party to cover another’s losses or expenses. It seeks unspecified damages and legal costs.

Criminal charges and parallel civil cases

Bloodworth faces 100 counts related to possessing sexually explicit material involving a child and 100 counts related to unlawful creation or distribution of deepfake visual material. He has pleaded not guilty, was released on bond, and was scheduled to return to court on September 21.

The cases turn on different questions

• Police said much of roughly 1,700 images and videos recovered from Bloodworth’s phone was created with Grok and photographs of juvenile clients. That is a police account, not a judicial finding.
• Several families have sued xAI over Bloodworth’s alleged use of Grok to generate sexually explicit images of their children. Those claims remain unresolved.
• Arkansas Act 827 criminalizes creating pornographic AI images of people without their consent, and lawmakers added AI-specific language to the state’s child sexual abuse material law.