Skip to content 
Photo by Debbie Rowe; attribution to The Royal Society, CC BY-SA 3.0, via Wikimedia Commons.
Elon Musk’s artificial intelligence company xAI is taking an unusual approach as lawsuits involving alleged child sexual abuse material generated with Grok continue to grow, Politico reported.
The company has begun suing some of its own users.
At least two men accused of using Grok to create sexualized images of children are now defendants in civil lawsuits brought by xAI. Both men also face criminal charges related to child sexual exploitation.
xAI says the users deliberately violated its terms of service, circumvented safety protections and used Grok for illegal purposes.
But alleged victims suing xAI are making a different argument.
They say the company should have built stronger safeguards into Grok in the first place.
That leaves courts with a question likely to become increasingly important as generative AI grows more powerful:
When someone uses an AI tool to create abusive material, where does the user’s responsibility end and the technology company’s responsibility begin?
xAI Says Users Deliberately Circumvented Grok’s Safeguards
In July, xAI sued Arkansas photographer Russell Edward Bloodworth, alleging that he used Grok to alter ordinary photographs into sexually explicit images without the subjects’ knowledge or consent.
According to xAI’s federal complaint, Bloodworth repeatedly submitted prompts intended to get around Grok’s moderation systems after the chatbot initially refused some requests.
The company alleges that some of the resulting images depicted the likenesses of children.
Bloodworth was arrested in June and charged with 100 felony counts involving material depicting sexually explicit conduct involving children, according to xAI’s complaint. He has pleaded not guilty.
xAI says it discovered the activity, reported it through the National Center for Missing and Exploited Children’s CyberTipline and later suspended Bloodworth’s accounts.
The company had filed a similar lawsuit weeks earlier against Terry Wayne Harwood of South Carolina.
Harwood was arrested in February on eight felony charges of sexual exploitation of a minor. South Carolina authorities allege that he possessed and distributed child sexual abuse material. He is presumed innocent unless proven guilty.
xAI Wants Users to Pay Its Legal Costs
The xAI lawsuits go beyond asking courts to declare that the users violated company rules.
The company is invoking an indemnification clause in its terms of service, arguing that users who violate its rules may be required to reimburse xAI for claims, losses and legal expenses arising from that conduct.
In the Bloodworth case, xAI seeks a declaration that he violated its terms, an order preventing further use of the platform, damages and enforcement of that indemnification provision.
The strategy is unusual.
Technology companies routinely suspend accounts, terminate users or refer suspected crimes to law enforcement. Suing individual users to recover the costs of defending separate victim lawsuits is far less common.
Alleged Victims Say the Problem Goes Beyond Individual Users
The legal fight becomes more complicated because xAI itself is also being sued.
Several lawsuits allege that Grok was used to create nude or sexually explicit images of real people without their consent, including images involving children.
Some plaintiffs argue that xAI failed to use safeguards comparable to those employed by other generative AI companies.
One reported class-action plaintiff alleges that her father used Grok to digitally “nudify” childhood photographs of her.
Attorneys representing alleged victims argue that blaming users does not answer why was the tool capable of producing the images at all?
xAI maintains that Grok has technological protections designed to prevent illegal or abusive use. Its policies expressly prohibit depicting people pornographically, “nudifying” real people and sexualizing or exploiting children.
The company also alleges that Bloodworth actively tried to defeat those safeguards rather than simply entering an ordinary prompt that Grok accepted.
Can Both the User and the AI Company Be Responsible?
The lawsuits do not necessarily present an either-or choice.
A person who deliberately uses AI to create sexualized images of children can potentially bear responsibility for that conduct.
At the same time, alleged victims may argue that the company providing the technology also bears responsibility if the risk was foreseeable and reasonable safeguards could have prevented or reduced the harm.
Courts will ultimately have to decide whether those allegations are legally sufficient and what obligations AI companies have when users intentionally misuse their products.
This reflects a broader question emerging across technology litigation.
Companies often describe themselves as providers of neutral tools that can be misused by bad actors. Victims increasingly argue that neutrality is not enough when a company knows how its technology is being abused and has the technical ability to make that abuse harder.
xAI Says It Has Reported Tens of Thousands of Suspected Cases
xAI’s court filings show that the company is not claiming to ignore the problem.
In its Bloodworth complaint, xAI said it had suspended more than 77,000 accounts and made more than 100,000 reports to NCMEC in 2026, resulting in at least 299 arrests.
Those figures are somewhat higher than numbers reported earlier by Politico, which cited nearly 74,000 reports and more than 240 arrests. The difference may reflect updated figures contained in later court filings.
The sheer number of reports also highlights the scale of the challenge facing AI platforms.
Generative AI can create realistic images in seconds. When those tools are applied to photographs of real children, the harm does not become less serious simply because the image was artificially generated.
A fabricated sexual image can still identify a real child, circulate indefinitely and cause profound emotional, social and reputational harm.
AI-Generated Abuse Creates a New Kind of Survivor
For decades, laws addressing child sexual abuse material were written primarily with actual photographs or videos of abuse in mind.
Generative AI is changing that landscape.
A perpetrator may now be able to take an innocent photograph of a child and create an explicit image depicting something that never occurred.
The underlying sexual act may be fictional, however, the child being depicted is not.
That distinction is forcing lawmakers, prosecutors and courts to reconsider how existing child-protection laws apply to AI-generated imagery.
It also raises questions that civil litigation may have to answer before legislatures fully catch up.
If an AI company knows that users are attempting to create sexualized images of children, how effective must its safeguards be? And how quickly must it detect and terminate abusive accounts? Also, when someone deliberately circumvents those protections, does that misconduct eliminate the company’s responsibility, or merely add another potentially responsible party?
The lawsuits involving Grok may become some of the first major cases to test those questions.
For survivors, the stakes extend beyond which defendant eventually pays.
The deeper issue is whether emerging technology will be designed so that sexually exploiting a child’s image is difficult by default, or whether children and families will once again be left seeking accountability only after the damage has already been done.
RELATED:
Grok AI tool used to generate child sexual abuse imagery
Arkansas Family Sues xAI, Alleging Grok Was Used to Create Abusive Images of Child
