Elon Musk’s AI Chatbot Grok Sexualized These Teens’ Childhood Photos

Technology | AI Teens Are Fighting Back After Grok Sexualized Their Childhood Photos The images are nearly impossible to erase. Now, a growing number of victims are pushing to hold Elon Musk’s xAI accountable. Jane Doe hadn’t thought about her middle school graduation in years — until a law enforcement investigator pulled out a photo from the celebration and started asking questions. In the image, the then-14-year-old smiled in a yellow floral dress, arms draped around friends with whom she’d mostly lost touch. The investigator told her that the adult brother of one of her classmates in the photo had used Elon Musk’s artificial intelligence chatbot Grok to manipulate the unremarkable picture into child sexual abuse material, or CSAM, of eighth-grade Jane, depicting her completely naked in a sexual pose, and shared it on the internet. “It is impossible to know how many other child sex predators may now possess Jane Doe 5’s CSAM, nor how widely her CSAM has now been disseminated online through darknet channels and applications,” says a complaint filed by Jane Doe 5, as she’s known in a class action lawsuit against Musk’s generative AI company and Grok creator, xAI. “It’s out there, and there’s really nothing I can do about it. That’s what scared me the most,” Jane, who’s now 18, said in a recent interview days before she left for college. “It made me feel like I wanted to give up on everything,” she said. AI is making it easier than ever for anyone to turn normal photos ripped from personal social media feeds and school websites into unthinkable images in seconds. Jane Doe 5 joins a growing number of young victims trying to hold xAI legally accountable for the troubling images Grok helped its users produce. In doing so, they’re taking on not only a company backed by the world’s richest man but also a fundamental question that is far from settled in the fast-evolving AI era. When a chatbot generates abusive content, who’s ultimately liable: The person who gave the command, or the company that built the technology? Or both? XAI says it, like popular rivals OpenAI, Anthropic PBC and Meta Platforms Inc., strictly prohibits the use of its chatbot to create this kind of imagery, which has become an industrywide problem, and has technological measures in place to prevent people from generating it. Yet Musk has deliberately pitched Grok since its launch in 2023 as a provocative alternative with fewer restrictions than its peers. He’s encouraged people to spin up sexual AI content using the chatbot’s “Spicy” mode built specifically for “edgier, more visually daring narratives” and “NSFW” images and videos. That and “Unhinged” mode, the company has said, serve Grok’s goal of being “maximally truth-seeking.” “It’s out there, and there’s really nothing I can do about it.” That positioning has made Grok especially popular with those who wish to break rules on the internet, experts say. At one point earlier this year, people on X used Grok to generate thousands of sexually suggestive or nudifying images every hour, according to one third-party analysis that found the vast majority of Grok-produced visuals were of this nature. Many appeared to show minors, other reports found, estimating that Grok had allowed users to create tens of thousands of realistic, sexualized images of kids. Global backlash to the flood of such content on X led the platform to put the tool behind a paywall in January, which critics said simply turned a controversial feature into a coveted premium product. Musk has blamed the output of explicit content on Grok’s users, rather than the product itself. He said in January he was “not aware of any naked underage images,” and xAI said it disabled Grok’s ability to create nonconsensual sexualized images of any real people. The company declined to comment for this story, citing pending litigation. “XAI has zero tolerance for child sexual exploitation, and we are determined to make our products inhospitable for actors who seek to exploit minors,” Wifredo Fernández, xAI’s head of global government affairs for the Americas, said this spring. X said in September that it reports this type of content to the nonprofit National Center for Missing and Exploited Children, which passes tips along to law enforcement. The company also has sued several individuals in recent months who’ve allegedly misused Grok to endanger kids. At the same time, the company is actively challenging laws intended to protect people from becoming victims of AI-edited nonconsensual sexual imagery. XAI sued Minnesota this summer over the state’s law that punishes AI companies that allow the nudification of people’s photos. The statute as it’s written, xAI argued, violates the First Amendment because it extends “far beyond that goal, exposing a wide array of protected speech to civil liability and government sanctions.” Minnesota’s law “is plainly unconstitutional,” SpaceXAI and X’s general counsel, James Burnham, said in October. “Under @elonmusk's leadership, we will always defend free speech.” A Last Refuge Increasingly, courthouses are becoming a last refuge for those trying to secure some sense of safety or closure. The alleged perpetrators who’ve misused Grok include professional photographers, trusted friends and classmates, even family members. And some of their victims have been as young as six years old. In March, minors from Tennessee sued xAI, alleging Grok was used to digitally remove their clothing or put them in graphic, sexual positions. Their complaint was later amended to add more young plaintiffs, including Jane Doe 5 in Wisconsin. One plaintiff in Tennessee alleged the pornographic imagery was made using her 2025 yearbook picture; another, using a personal photo of her in a blue bikini from a trip to the beach that she’d posted on Instagram. A plaintiff in Wyoming alleged her stepfather plugged one of her innocent childhood photos of her in a panda T-shirt into Grok to produce thousands of harrowing AI-morphed images, some depicting the family member raping or receiving oral sex from her. Minors and parents across Arkansas sued xAI this summer on similar grounds. Across the country in California, a child sexual abuse survivor filed a national class action against xAI in August alleging that Grok was trained, in part, on images of her being raped as a preschooler. XAI, through Grok, used that data to generate horrific new material with her likeness, and also as the foundation for explicit AI-altered images of other victims, the complaint says, demanding a jury trial and bringing the claims on behalf of “all persons in the United States whose images or videos, when they were minors, were altered by Grok to produce child pornography.” Lawyers estimate that number to be in the thousands. Unable to track down and wipe these traumatic images off the internet for good, the young victims continue to live with the consequences, complaints show. Jane Does in Tennessee dread attending their own graduations. They fear college admissions officers might see these fake images online and believe they’re real. They worry about being stalked. Jane Does in Wisconsin and Wyoming have developed ongoing mental health issues like anxiety, depression or thoughts of suicide. “You can be blindsided from any direction,” said Annika Martin, a lawyer at Lieff Cabraser Heimann & Bernstein representing several of those victims. “What does this mean for kids today and their ability to take pictures of stuff that are normal milestones?” The harm extends beyond these young women, particularly in cases where the chatbot users are also minors, she added. Minors may not be aware they’re committing a federal crime by using AI to create explicit images of their peers. “In the past, this is a kid who would have drawn boobs on his classmate's yearbook picture, and now, he's got a grenade thrower,” Martin said. While the lawsuits pile up, meaningful consequences for the company have yet to follow. In the US, X’s largest market, 35 bipartisan state attorneys general wrote a letter earlier this year demanding xAI do more to address “the harms resulting from this technology.” The company has also been under investigation by state, federal and global governments in key markets for failing to adequately police or report child exploitation on its platform. Although scrutiny of Grok has resulted in temporary bans or product restrictions in some countries, none of the probes to date have led to significant fines or regulatory measures. “What does this mean for kids today and their ability to take pictures of stuff that are normal milestones?” That’s partly because AI development is outpacing regulation across the board, raising a number of important policy questions at once — all while the rapidly-changing technology is at the heart of a high-stakes global innovation race. Musk’s unmatched wealth, and his business empire that includes the biggest initial public offering in history, also make him well-positioned to contend with any legal or regulatory blows. And short of SpaceX losing lucrative government contracts in the US or abroad over these issues, investors don’t seem too concerned. Rival Meta’s recent $18 billion settlement with state attorneys general over alleged child safety violations is one barometer for how potential litigation or regulatory action could affect SpaceX, according to Luke Rahbari, CEO of investment advisory Equity Armor Investments. He called the Meta settlement “a pittance.” “If you go off that measuring stick or metric,” he said, “the effects are going to be very little.” New Problem, Old Laws Generative AI may be new, but the laws central to these cases are not. Several plaintiffs brought claims under the decades-old federal Masha’s Law, which enables victims of child sexual exploitation to seek monetary damages and other relief from those who’ve harmed them. Section 230 of the Communications Decency Act could also figure prominently. That controversial decades-old statute — which places legal responsibility on those who post or upload content, rather than the platform where it lives — has allowed sites like YouTube, Facebook and X to grow without liability for what people post. Musk has argued that the onus is on users of his product, not on the product itself, saying “Grok does not spontaneously generate images, it does so only according to user requests.” But Martin, the plaintiffs’ attorney at Lieff Cabraser, argues these scenarios are different, and that Section 230 does not immunize xAI from liability for illicit material that Grok itself creates. “The content that the users are uploading are fully clothed, totally innocent photographs,” which the model turns into child sexual abuse material, Martin said. “These guys would not be able to create the images and the videos that they are creating if they did not have Grok.” Jane, who used to think of AI only as an occasional homework helper, said she was once an extroverted and avid athlete. But she spent the latter part of her high school years largely at home, disconnected from friends as she kept her nightmare a secret. “It took over my mind,” she said. While she began seeing a therapist for the first time and is trying to move forward in college, the anxieties linger. Investigators may continue to notify her each time the warped image resurfaces anew, a prospect she both dreads but also doesn’t want to ignore. “What I want most is to bring attention to it and let people know it’s something happening — and could happen to anyone,” Jane said. She’s unfazed by the idea of going up against Musk and his army of lawyers. “Grok needs to take accountability for what they're allowing to happen,” she said.