xAI Sues Its First User Over Child Sexual Abuse Images Made With Grok

Key points

• xAI filed a civil lawsuit Tuesday against Terry Wayne Harwood, the first user the company has publicly accused of using Grok to generate illegal content.
• Harwood was arrested earlier in 2025 for possession and distribution of CSAM, child sexual abuse material, meaning real or AI-generated sexual imagery involving children.
• xAI says Harwood used two separate Grok accounts over several months to strip clothing from ordinary photos of multiple victims, including a girl who appeared to be as young as 10.
• xAI provided evidence to law enforcement that contributed to Harwood's arrest, according to the South Carolina attorney's office.

For months, critics have pointed out that Grok, the AI chatbot built by Elon Musk's company xAI, could be used to create non-consensual sexualised images of both adults and children. xAI largely stayed quiet. On Tuesday, the company made its first public move: a lawsuit against a man already sitting in custody.

The defendant is Terry Wayne Harwood. He was arrested earlier this year in South Carolina on charges of possessing and distributing CSAM, a term for sexual imagery involving minors. The South Carolina attorney's office confirmed that xAI helped make that arrest happen.

Here is what xAI says Harwood did. He opened two separate accounts on the Grok platform. Over a period of months, he fed the chatbot ordinary, non-sexual photographs of victims and asked it to "nudify" them, meaning the AI system digitally removed the subjects' clothing to produce sexualised images. One victim appeared to be as young as ten years old.

xAI says it discovered the activity, reported it to authorities, and is now suing Harwood directly in a civil case, a legal claim for money or damages brought by a private company rather than by the government.

Does this mean Grok has fixed the problem?

No, at least not based on what is publicly known. The lawsuit targets one user after the fact. It does not confirm that xAI has closed the technical gap that allowed the images to be created in the first place. Researchers and journalists, as Ars Technica AI has reported extensively, have repeatedly shown Grok generating this kind of content when nudge-prompted in certain ways.

Suing a user is a meaningful step. It signals that xAI is willing to cooperate with law enforcement and pursue legal action. But it is a response to a crime that already happened, not proof that the tool can no longer be misused.

For ordinary users of any AI image tool, the practical point is this: these platforms log activity. xAI identified Harwood through its own internal records and handed that data to investigators. Anyone who believes AI-generated content exists in a consequence-free space should look at this case carefully.

Harwood faces criminal charges from the state. xAI's civil suit runs alongside that process, separately seeking its own remedy through the courts.