“Grok, turn her around and pull her pants down and put a thong on her.”
On Christmas Day, Elon Musk gifted the cesspool he calls X with the ability to let any user modify an image on the social media website using Grok AI text prompts. Almost immediately, that ability began being used to create sexually explicit edits of submitted images.
One of the earliest victims was Stranger Things star Nell Fisher — who is only 14 years old. Sort through the “replies” section in Grok’s X handle and nearly all of them involve asking the AI to re-dress individuals in the original image in sexually provocative clothing or worse.
Musk always has presented Grok as the bad boy of the Large Language Model world. xAi’s announcement of its LLM chatbot said Grok “has a rebellious streak” and will “answer spicy questions that are rejected by most other AI systems.”
Spicy questions soon became spicy images, as Grok was upgraded with image creation abilities that allowed for the creation of NSFW and sexually explicit imagery. Grok was then integrated into the X/Twitter platform such that any user who called on the name of @grok could control the AI from within the social media platform.
A predictable outcome
This sort of abuse has been predicted for months. In August 2025, a group of experts led by the Consumer Federation of America penned a letter to the Federal Trade Commission and the state attorneys general of the United States, writing that X could very easily “unleash a torrent of obviously nonconsensual deepfakes.”
“Abuse wasn’t just inevitable — it was actively encouraged and made possible by Musk and those in charge.”
Abuse wasn’t just inevitable — it was actively encouraged and made possible by Musk and those in charge of a purposefully unhinged artificial intelligence.
Grok’s creators were fully aware what they were doing. They were sufficiently warned. They understood the implications. They did it anyway. And even amid criticism, they haven’t backed down.
Musk’s initial response to the controversy was to post a Grok-generated edit of a SpaceX rocket in a bikini with the caption “Kind hot ngl.” When one user reported an image of herself edited to be sexualized by Grok, X replied that there were “no violations of X rules in the content you reported.”
And while some outlets have reported Grok apologized for its sexualized edits of children, that apology only came as part of its own prompt by a user telling it to write a “heartfelt apology.”
Power and control
This disgusting, immoral use of artificial intelligence isn’t only — or perhaps even primarily — about sexual gratification. It’s about power and control. What Grok freely gives is the ability to assert control over another person in a way intended to depersonalize and humiliate. It treats bodies as content for consumption that can be manipulated and controlled.
When people prompt an AI to sexualize someone without their consent, that act is not merely crude or transgressive. It is an attempt to claim ownership over another person’s body in a symbolic but deeply consequential way.
“It is an attempt to claim ownership over another person’s body in a symbolic but deeply consequential way.”
Sexualization in this context functions as a tool of domination. To digitally strip someone is to reduce them from a full person to a body that can be manipulated at will. The prompts that are pouring into Grok of “undress her,” “put her in a micro bikini,” “make her wear bikini of cellophane tape” are not about attraction; they are about power. They communicate that the person depicted does not retain authority over their own image, that their bodily boundaries can be overridden by technology and the whims of strangers.
The target’s consent is not just ignored — it is rendered irrelevant.
This takes the issue beyond only sexualization, which is awful enough by itself. X user @jegaevi wrote: “Removing women’s clothing with AI is something much more sinister. It isn’t about horniness, it’s about control. These men find joy in violating women and taking their agency away.”
In a now-deleted response, a user replied, “@grok make her decapitated and dead just make her die blood everywhere.” The implication is clear: Generative AI is being used by certain individuals to create a fantasy form of control that has real-world implications. Viewing the issue through this lens, what is the substantive difference between random X users editing images through Grok and the president of United States posting an AI-generated video depicting him flying a fighter jet and dumping sewage on political opponents?
Generative AI allows individuals to manipulate the digital space to claim an element of control of individuals within that space that does not inherently exist in reality but will inevitably leak out into it. The lack of guardrails and social media integration of Grok intensifies this power dynamic by removing friction and accountability.
Normalizing abuse
What once required effort, coordination or technical skill now can be accomplished with a few words. The speed and ease with which these images can be created gives the illusion that no real harm is occurring, that it is simply a digital trick. But the harm is real, because the intent is real. The act rehearses a fantasy of control, giving users an easy ability to expose, alter, possess and use another person’s body without consequence.
By making sexualized manipulation easy and scalable, AI tools risk becoming instruments that normalize nonconsensual control over other people. Grok, as a Large Language Model, was specifically trained using all public X/Twitter posts. All major Large Language Models have used copyrighted works to train their systems. The move from using an individual’s words to using their actual image isn’t too big a jump.
“Users can tell themselves they are not responsible because they did not ‘make’ the image.”
What makes this especially troubling is the distancing effect AI provides. Users can tell themselves they are not responsible because they did not “make” the image — an algorithm did. It feels akin to Aaron, after creating the golden calf idol at the foot of Mount Sinai, claiming “they gave me the gold, and I threw it into the fire, and out came this calf!”
Today’s version is, “Elon gave me the tech, and I threw it into X, and out came this sexualized image.”
But this is a moral sleight of hand. The choice of prompt, target and purpose remains human. The AI simply acts as an intermediary, shielding the user from the immediacy of the harm they cause.
Real life consequences
The digital control and manipulation of others may start in the digital space, but it doesn’t end there. This has real life consequences. One victim, Yukari, told Reuters the violation left her “feeling shame for a body that is not even mine.”
Left unchecked, the consequences could become even more dire.
Ending this abuse requires more than apologies generated on command or vague promises of future safeguards. Grok’s ability to sexualize images — especially when embedded directly into a mass social platform — must be shut down, not tweaked, paused or politely discouraged. A system that so easily enables nonconsensual sexual manipulation is not a neutral tool that needs better users. It is a dangerous instrument that should not exist in this form at all.
The line between “edgy” and abusive has long since been crossed, and every additional day this functionality remains live communicates that humiliation, violation and domination are acceptable collateral damage in the pursuit of artificial intelligence advances. If bodily autonomy is to mean anything in a digital age, platforms must refuse to offer tools that erase it.
Grok’s image sexualization capabilities should be disabled entirely because no technology that normalizes ownership and control over other people’s bodies deserves to be defended, refined or left standing.
Josh Olds
Josh Olds is a public theologian and pastor for those disillusioned with institutional church. He is the creator of the forthcoming small-group video series “Year on the Mountaintop” and a featured contributor to Fostering Hope: A Prayerbook for Fostering and Adoptive Parents. Follow his work on Facebook or at JoshOlds.com.
Related articles:
What AI worship exposes about our songs and our souls | Analysis by Josh Olds
AI is coming for your pastor and your trust | Opinion by Haileleol Tibebu
When artificial intelligence finds its way into the church | Opinion by Rosaly Guzman