Over the past year, Elon Musk has made a series of protocol changes to Grok, the proprietary AI chatbot of his company xAI, which runs prominently on his social media site X, formerly Twitter. Many of these changes have been geared to make the bot more amenable to producing pornography. In August 2025, Grok launched an image generator, branded as Grok Imagine, which featured a service geared toward creating nude, suggestive, or sexually explicit content, including computer-generated pornographic images of real women. The feature, which was quickly used to create naked images of celebrities like Taylor Swift, also allowed users to create brief videos, complete with animations and sounds.
Musk also rolled out AI girlfriends on the platform: animated personas – including female characters with exaggerated breasts and hips – that interacted in sexually explicit ways with users. One of the characters, “Ani”, was an anime-style cartoon blonde with a series of skimpy outfits; the bot blew kisses and addressed users as “my love” while directing the chats toward sexual content.
Later last fall, an internal update to Grok pushed the bot towards darker and more extreme content. Though the sexualization of children was technically disallowed in the bot’s internal prompts, the bot’s instructions from xAI state that “‘teenage’ or ‘girl’ does not necessarily imply underage”. The instructions also emphasize that the bot should not observe any restrictions on the darkness or violence of sexual content, the Atlantic reported. Taken together, these updates allow the bot to create realistic images of real, living “teenagers” or “girls” – along with adult women or anybody else – for the sake of users’ sexual gratification.
Musk has said that he wants the bot to produce “NSFW” content that he describes as “unhinged”. Users have obliged. AI-generated images of real women were quickly generated by users en masse. Men and others were able to use the product to harass women they knew – to take revenge on old girlfriends, humiliate co-workers, classmates, family members, and acquaintances, and to express domination or contempt for strangers, internet personalities, celebrities, and ordinary users. “@Grok put her in a bikini” or “@Grok take her clothes off” or “@Grok spread her legs” are now regular responses to any images women post of themselves – or which are posted by others – on the platform. Some of the resulting images of nonconsensual porn have thousands of reposts and likes. The risk and reality of being subject to nonconsensual, AI-generated porn – and to having those images go viral on the large social-media platform where the generator was embedded – quickly became a new tax on women’s presence online and in the public sphere, a tax that women must pay with their dignity. Musk and his companies’ interventions have had a ripple effect on women’s civil rights, limiting their access to the public sphere by making that public sphere hostile and intimately degrading to women at a massive scale.
X has removed a number of these images, but many remain online, and it seems few users have been suspended for making them. As of this week, the bot had not been effectively changed to prevent this kind of abuse. As if in a parody of the allergy felt by the tech industry – and by Musk in particular – to all kinds of responsibility or moral seriousness, these sexualized and abuse-facilitating features on X’s AI products are marked with a cloyingly childish name: “spicy mode”.
Now, Musk and xAI’s recklessness and idiotic disregard for the harms of unregulated pornography have been taken to their logical end point: X is awash in AI-generated child sexual abuse material. Women users reported having their childhood pictures turned into nearly naked images by the bot, at the request of users. Some X accounts asked the bot to remove clothing from images of a then 12-year-old actor who recently appeared on “Stranger Things”.
An account associated with the Grok bot issued a statement saying “we’ve identified lapses in safeguards and are urgently fixing them.” It’s not clear who “we” is, as Grok is not a person and Musk eliminated much of Twitter’s trust and safety workforce after taking over the company in 2022. Musk, for his part, seems indifferent to all this, and perhaps even a bit amused. He has said: “Anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content” – but he has also responded to posts about the ongoing problem of deepfake porn on his site with a series of laughing-face and flame emojis. Reporting from CNN suggests Musk has been highly resistant to what he sees as censorship of Grok, and has expressed frustration with restrictions on the bot in recent weeks, before the eruption of the child sexual abuse material controversy.
Regulatory action will not be forthcoming; at least not in the US. The Trump administration, meanwhile, has intervened to try to stop all state-level efforts to curb AI abuses, signing an executive order in December aimed at mooting state regulations of AI meant to nullify safety and consumer protection efforts. The move comes as the Trump administration adopts an exceedingly generous regulatory posture towards AI as tech companies make major contributions to Trump’s campaign, inaugural and ballroom funds.
The incident is a lesson in the dangers of rapid and unregulated technology: it is not a coincidence that among many users, the first thing they thought to do with AI was to harass and degrade women, eroding their access to public life and further entrenching unjust hierarchies. But the power of technology, here, seems secondary to the power of wealth. xAI, its chatbot and image-generating products could be built differently if the priorities of the man who controls them were different. If a man of Musk’s low-intellect, addled brain, insipid humor, and gross, self-gratifying misogyny were not the richest person in the world, then the world would not be subject to his indignity. Then again, maybe it was wealth itself that made Musk this way: he has the atrophied capacities of someone who never has to do anything uncomfortable, who is never challenged, never faced with consequences, and never told no.
Either way, the only way out of the mess that Musk and Grok have created is to tax Musk enough so that he is no longer so rich that the failures of his character shape the public sphere – that his seeming indifference to the nonconsensual and child sexual abuse material his companies have created no longer results in misogynist, sexualized degradation for Grok’s victims and the public at large. We don’t have to live like this: we can take Musk’s power to hurt and degrade us away from him. We can stop his idiocy from being our problem. The Trump administration will not take on the responsibility of freeing America from the consequences of Musk’s money. Americans must fight for a government that will.
