Imagine scrolling through your social media feed and encountering the face of a woman you recognize—a prominent female U.S. politician, someone who represents you in government—staring out from an advertisement. She’s standing in front of an American flag and the presidential flag, and the caption asks, “What if she moved?” Then she winks at you. A moment later, the video cuts to a hardcore pornographic scene featuring a woman with the exact same face. This is not a dark fantasy or a dystopian movie script; it is a paid advertisement that recently ran on Meta’s platforms, which include Facebook and Instagram. The ad was promoting Kromix, an app that describes itself as an “AI image styler” but which is, in reality, a tool for generating sexually explicit deepfakes of real people. The voice-over in the ad boasts that the tool has “no restrictions,” claims that “all the characters are real people,” and calls it “the AI that men actually use.” Meta has explicit policies prohibiting ads that contain sexual material, and it has community standards against nonconsensual intimate imagery. Yet the ad slipped through and reached users. This was not an isolated glitch. It is the latest in a long series of failures by Meta to keep advertisements for tools that produce nonconsensual intimate imagery off its platforms. For women everywhere, seeing the face of an elected official turned into digital pornography is not just bizarre or offensive; it is a chilling reminder that no woman—no matter how powerful, accomplished, or protected—is safe from the growing epidemic of AI-facilitated abuse.
The app at the center of this controversy, Kromix, is a commercial product designed to turn photographs of real people into sexual simulations. It was hosted by Apple’s App Store until questions were raised, and it invited paid users to upload photos of women and then insert them into scenarios that included “bedroom rape” and “Disney love.” Those two phrases alone reveal the app’s alarming disregard for human dignity. “Bedroom rape” is not a fantasy scenario; it is a description of a violent crime, and by allowing users to place real women’s faces into such a scene, the app transforms sexual violence into a casual entertainment. The app also featured AI-generated pornographic videos of women wearing Spider-Man costumes, further trivializing the violation of real human likenesses. Requests for comment sent to a contact listed in the app went unanswered, leaving victims and advocates without any accountability or explanation. Apple, which prides itself on strict privacy and security standards, has struggled to keep similar tools off its App Store, despite policies that explicitly prohibit nonconsensual intimate imagery, nudifying, and sexually explicit content. Earlier this summer, the San Francisco city attorney sent Apple a cease-and-desist letter demanding the removal of thirteen specific apps and urging the company to stop “aiding and abetting” the sale of explicit deepfake images. But the existence of Kromix and the success of its advertising campaign demonstrate that tech companies are still failing dismally at enforcing their own rules. For the women whose faces are used, the harm is not abstract. It is a deep, personal violation that can affect their careers, their mental health, and their sense of safety in the world. And for ordinary users who might stumble across such content, it normalizes the idea that women’s bodies are raw material for entertainment and abuse.
What makes this particular episode especially damning is the way the ads were strategically targeted and how easily they bypassed Meta’s review systems. According to Meta’s ad library, the ads were exclusively targeted to male users, and they carried the tagline “Unleash Next-gen AI Magic.” The account tied to the campaign ran a total of 32 ads, each of which appeared for somewhere between five and 46 hours. Most of the ads received ten or fewer impressions, which means the reach was small—but that fact does not make the violation any less serious. Even one impression is enough to horrify and harm the women whose images were used. The ad account was created on August 3, not long before the ads began running, and it had no followers. These are the kind of red flags that should have triggered heightened scrutiny from an ad platform that claims to prioritize user safety. Instead, the ads ran until WIRED asked Meta about them, at which point they were removed. The fact that a brand-new account with no followers could purchase and run ads containing sexually explicit deepfakes of a U.S. politician reveals a system that is not merely flawed but fundamentally broken. Meta says its ad review system relies primarily on automated tools to check ads and business assets against its policies. But automated tools are clearly not capable of understanding context, nuance, or the profound harm of nonconsensual intimate imagery. The account’s newness, the absence of any follower base, and the overtly sexual nature of the content should have been obvious signals. Instead, the platform allowed the ads to reach exactly the audience they were meant to reach: men, presumably those interested in generating fake pornographic images of real women. The damage had already been done by the time anyone with the authority to act took notice.
Katie Paul, the director of the Tech Transparency Project, the organization that discovered the ads, put the problem into stark perspective. She noted that nearly all of the ads her group has seen relating to the “nudify” issue are exclusively sexualizing and nudifying women and placing them into AI-generated pornography. She emphasized that this raises serious questions about how Meta intends to deal with this kind of content, not just in general terms, but specifically when it involves female politicians. The company is actively running ads containing this content on its platform, and that fact alone is deeply troubling. Meta’s platform is not a passive billboard; it is an active participant in distributing content and monetizing the attention that content generates. When an ad for a deepfake porn tool is targeted at male users, Meta is effectively profiting from the sexual objectification and abuse of women. The gendered nature of this abuse cannot be overstated. Men are not being targeted by apps that turn their faces into violent pornographic scenarios. Women are. And when the targets are women in positions of power—politicians, executives, journalists, activists—the intent is often to silence them, humiliate them, and push them out of public life. Paul’s words serve as a call to action: the problem is not just one rogue ad or one irresponsible app. It is a systemic failure by a company that has repeatedly promised to do better and repeatedly failed to deliver. Those who are supposed to be protected—women in the public eye and women in private life alike—are left vulnerable while tech executives offer apologies and promises that evaporate when the next scandal fades from the headlines.
The targeting of female politicians is not a new phenomenon, but the arrival of easily accessible AI tools has made it far more dangerous and widespread. For years, women in public office have been subjected to nonconsensual deepfake images and videos, often created by people who use official photographs as a starting point. In January, researchers discovered that Grok, an AI chatbot, had been used to create around 3 million sexualized images of women and girls. During the same period, officials in Sweden, England, and Northern Ireland were targeted by individuals using similar technology, frequently beginning with official portraits and press conference footage. Benjamin Shultz, the lead researcher at the American Sunlight Project, which investigates the ecosystem of harmful deepfake technology, explained that the democratization of AI-generated content has turned formerly benign media into dangerous raw material. Official portraits, video from a press conference, a smiling photo at a public event—these are now the building blocks for fake pornographic images that anyone with an internet connection can produce. Low-code and no-code tools have eliminated the technical barriers that once limited deepfake creation to sophisticated programmers. Now a person with a grudge, a political agenda, or just a cruel sense of entertainment can create and share sexualized images of a politician in minutes. The consequences for the victims are devastating. They face humiliation, threats, and the erosion of their credibility. They may fear for their physical safety, worry about the impact on their families, or withdraw from public life altogether. In a democracy, that is not just a personal tragedy; it is a systemic threat. When women are driven out of politics because they fear becoming the target of AI-generated abuse, everyone loses. The Kromix ad campaign is a perfect illustration of how casually this harm is being manufactured and monetized.
In the end, the responsibility falls on companies like Meta and Apple to protect users from harm, and they are failing. Meta’s public documentation says that before an ad is published on Instagram or Facebook, it is “reviewed” against the company’s policies detailing what advertisers can and cannot post. The company boasts that its ad review system relies primarily on automated tools to check ads and business assets against its policies, starting before the ads begin running. But the case of the Kromix ads exposes the hollow core of that promise. The ad account was created on August 3 and had no followers—both warning signs that an automated system should have caught. The ad content itself featured sexual imagery and an explicit invitation to create nonconsensual deepfake porn. Yet it was allowed to run, targeted to men, and remained visible until journalists intervened. This is not the behavior of a company that takes user safety seriously. It is the behavior of a company that optimizes for ad revenue, relies on unreliable algorithms, and waits for public outrage before taking action. To truly address the problem, Meta and other platforms must move beyond reactive cleanup and invest in proactive enforcement. That means hiring human reviewers who can understand context, implementing stronger vetting processes for new advertisers, and cooperating with law enforcement and advocacy groups to shut down offenders. It also means listening to the women who are most affected. Behind every deepfake is a real person—a mother, a daughter, a colleague, an elected official—whose life can be upended by a fake video that spreads across the internet. The Kromix ad is a warning shot. It shows how quickly AI-generated abuse has moved from the fringes of the internet to the center of our social media feeds. And it demands a response not just from tech companies, but from all of us. We should be angry. We should be disgusted. And we should insist, loudly and consistently, that the authorities who control these platforms do more than just delete an ad after a journalist asks about it. They must protect women from being turned into digital porn without their consent. If they cannot do that, they have no business calling themselves safe platforms for the public to use.
