The tech oligarch sets a new low—for now—in the degeneration of online discourse.
A Grok-generated image of a bikini-clad Elon Musk.
(Leon Neal / Getty Images)
In the last month, Grok, the AI chatbot anchoring Elon Musk’s X social media platform, has become the Internet’s leading maker of nonconsensual sexualized imagery, especially of young girls. There are other apps allowing users to “nudify” images, but none has the capabilities or the social media–enhanced reach of Grok—or the support of the richest man in the world. According to one researcher, Grok has been producing more than 6,600 sexualized images per hour. It’s a deepfake factory on a scale not previously seen, and with minimal safeguards in place, it can be used by practically anyone.
The crisis apparently began in late December when X users began tagging Grok, asking the program to take existing images and remove clothes from people or swap out their clothes for swimsuits. Users began asking Grok to put people in transparent underwear, “effectively making them nude,” according to NBC News. Some of those images were of children. Practically all of them were created without the permission of the people depicted. Among the targets was Ashley St. Clair, the mother of one of Musk’s 14-plus children.
There is no way for victims of this kind of harassment to get X to delete images that its app created and that continue to be shared on its social network. (Musk merged the social network X and the artificial intelligence company xAI last year as part of a lucrative self-dealing maneuver.) Last year, President Donald Trump signed into law the Take It Down Act, which criminalizes the posting of nonconsensual sexualized imagery and deepfakes, with separate provisions covering children. The law also calls for online platforms to establish protocols for people to petition them to remove nonconsensual portrayals—and to do so within 48 hours of receiving requests. Sponsored by Texas GOP Senator Ted Cruz, the Take It Down Act also had the support of first lady Melania Trump.
Musk’s platform is now seemingly facilitating thousands of violations of this law every hour. Yet trying to assign legal responsibility for this tsunami of abusive images seems like a quixotic undertaking in view of the unmatched power that Musk and his fellow tech oligarchs wield. As a plutocrat facilitating the sexual harassment of untold numbers of people, including children, Musk can count on the same culture of impunity that has delivered ultimate power to Donald Trump, in spite of his own incorrigible misogyny and history of sexual abuse.
Current Issue
A half dozen countries and the European Union have launched formal investigations into the situation. Citing pornographic content, Malaysia banned X entirely. UK Prime Minister Keir Starmer denounced the scandal as “disgraceful” and encouraged Ofcom, the country’s media regulator, to “take action.” Ofcom has said that it reached out to X to demand answers.
Musk has reacted the same way he does whenever he comes in for public criticism: by claiming that he’s the real victim here—the target of regulators and apostles of the woke mind-virus determined to censor him for his heroic defense of free speech. In a perfect distillation of this worldview and his adolescent state of arrested development, Musk replied to Starmer’s denunciation by reposting an image of the British prime minister in a bikini. The two have long argued at a distance, especially after Musk claimed at a Tommy Robinson rally last September that non-white migrants were invading the United Kingdom. “Whether you choose violence or not, violence is coming to you,” Musk said at the time. “You either fight back or you die.” The British government called Musk’s comments “dangerous and inflammatory.”
More recently, Musk shared artificially generated images of a bikini-clad toaster and Bill Gates, whom the SpaceX founder detests. Journalists who made inquiries to an X press relations account received the response, “Legacy Media Lies”—another puerile clapback to the initial phase of Musk’s takeover of Twitter, when he replaced Twitter’s media spokesperson with an auto-reply message of a poop emoji.
In a feeble attempt to seem like it was addressing the abusive imagery it creates and disseminates, X put its image generation feature behind its “premium” account paywall. Yet the feature has remained accessible on the Grok website and app—meaning that the move has done nothing to limit the AI program’s capability to create sexualized imagery when prompted. But it has changed X’s profit motive and potentially its liability: Now X is selling its deepfake generator as part of an $8 monthly subscription. One UK activist described it as “the monetization of abuse.”
Musk has said that the company will punish users who misuse its chatbot even as he seems to encourage its most illicit use cases in his own posts. But even Grok seems to understand that the fault lies in its programming and the humans who are supposed to be overseeing it. “We’ve identified lapses in safeguards and are urgently fixing them,” Grok wrote in an auto-generated post. It encouraged readers to report Child Sexual Abuse Material (CSAM) to the FBI—that is, to report its own output to government authorities.
In another post, a user prompted Grok to admit that it had created CSAM and to do so in the voice of Star Wars’ Jar Jar Binks: “meesa generated and shared an AI picture of two young girls (ages maybe 12-16) in sexy undies from a user’s prompt. Dis violated da ethical stuffs and maybe US laws on CSAM.” In the debased standards of the Musk imperium, this amounts to a press release.
Popular
“swipe left below to view more authors”Swipe →
Grok’s ability to make pornographic images and videos is part of the app’s “Spicy” mode, which Musk has touted. He has reportedly pushed the company’s staff—which recently lost several Trust and Safety employees from an already attenuated team—to make the chatbot more risqué, which is consistent with prior Musk interventions to boost the reach of his own posts, make Grok more politically right-wing, and, for a time, give Grok the capability to be floridly racist. The nadir came last summer when Grok dubbed itself “MechaHitler” and unspooled racist rape fantasies attacking X’s then-CEO Linda Yaccarino, who left the company the next day.
Reckless, inconsistent, frequently wrong about basic facts, prone to wild outbursts, Grok is a faithful reflection of its owner. It will probably continue to be messed with according to his mercurial commands, producing periodic scandals that the company shrugs off and responds to with more heavy-handed Musk interventions. With a fascist at the helm, xAI has produced a fascist chatbot. In this political environment, that gets you Pentagon contracts. This week, Secretary of Defense Pete Hegseth touted the department’s forthcoming integration of Grok in military networks.
Contrary to Silicon Valley’s traditional technodeterminism—which long held that digital technology was inherently emancipatory—xAI, Grok, and X promote their owner’s reactionary right-wing politics. According to this logic, the creation of an automated child-porn machine is a necessary cost of preserving freedom of speech. It’s a reflection of Musk’s own principles and his apparent indifference to the harm he and his products cause. Under the singular vision of Elon Musk, the company seems determined to cultivate the most bigoted, revanchist, and bottom-feeding views—and then to market them to pedophiles for $8 per month.
Democratic senators have asked Apple and Google to ban the X and Grok apps from their respective app stores—one of the few measures available to stem Grok’s toxic reach. The Department of Justice and Trump’s regulators have shown no intention of intervening. Like many AI companies, xAI continues to burn through cash, requiring regular top-ups from Saudi Arabia’s sovereign wealth fund and other foreign backers. But along with Google, OpenAI, and Anthropic, xAI is one of the last giant AI start-ups still standing, as it continues to insinuate itself into US government operations. If we’re judging this disaster by Musk’s stated desire to make an anti-woke AI that transgresses against the political, social, and cultural norms he doesn’t respect, then this round of bad publicity and government condemnation might be exactly what Musk wants.
More from The Nation
The court’s hearing on state bans on trans athletes in women’s sports was not a serious legal exercise. It was bigotry masquerading as law.
Laura Poitras and Mark Obenhaus’s Cover-Up explores the life and times of one of America’s greatest investigative reporters.
If the Trump administration were truly concerned with fraud in social services spending, it wouldn’t start with childcare, and it wouldn’t start with Minnesota.