Meta’s ad systems again came under pressure after ads appeared on its platforms for Kromix, an AI tool promoted in ways that appeared to invite users to create deepfaked pornographic videos resembling female US politicians.
The campaign was short-lived, but the details matter. The ads ran despite Meta’s policies against sexual material in advertising and despite broader rules barring nonconsensual intimate imagery and promotions for nudify or AI kissing apps.
What the Kromix ads showed
Kromix presents itself as an “AI image styler.” According to the source article, one video ad viewed by WIRED used a voice-over that described the tool as having “no restrictions,” said “all the characters are real people,” and called it “the AI that men actually use.”
One ad showed a woman closely resembling a prominent female US politician in front of a US flag and the flag of the president of the US. The caption read “What if she moved?” The ad then shifted to a pornographic video featuring a woman with the same face.
The app itself, which had been hosted on Apple’s App Store before WIRED requested comment, invited paid users to upload photos for scenarios including “bedroom rape” and “Disney love.” The source article also says videos on the app featured AI-generated pornographic videos of women wearing Spider-Man costumes.
Requests for comment sent to a contact listed in the app went unanswered.
How the campaign reached Meta’s platforms
Meta’s ad library showed that the ads were targeted exclusively to male users. The campaign used the tagline “Unleash Next-gen AI Magic.”
Meta said the account tied to the campaign ran 32 ads. Those ads ran for between five and 46 hours, and most received 10 or fewer impressions. The ads were removed after WIRED asked Meta about them.
The page behind the ads had been created on August 3 and had no followers. In the source article, those details are described as possible warning signs that the account might have been suspicious.
The case also shows one method advertisers used to obscure what the ads contained. Most of the video ads used benign-looking cover photos, such as a beach, mountains, or a waterfall, while the rest of the content was pornographic.
Why female politicians are a major target
The Kromix ads sit inside a broader problem: nonconsensual deepfake images and videos have targeted politicians and elected officials for years, and the targets have been overwhelmingly women.
In January, Grok was used to create around 3 million sexualized images of women and girls. Officials in Sweden, England, and Northern Ireland were among those targeted by people using the technology, often starting from official images.
Benjamin Shultz, the lead researcher at the American Sunlight Project, is investigating how harmful deepfake technology is used against politicians. His point, as described in the source article, is that once AI content production becomes easier to access, ordinary public materials can become inputs for abuse.
That is the central risk for public figures. A press image, official portrait, or public video may be harmless in its original setting, but low- and no-code tools can turn the same material into nonconsensual intimate imagery.
Meta’s rules and the enforcement gap
Meta’s public documentation says ads on its platforms, including Instagram and Facebook, are reviewed against its policies before publication. Its advertising standards say the review process relies primarily on automated tools to check ads and related business assets.
The company’s ad rules bar sexual imagery, including nudity, sexual activity, and sexually suggestive content. Meta’s broader adult sexual exploitation policies also prohibit nonconsensual intimate imagery and the promotion of nudify or AI kissing apps.
Yet ads for nudify and undressing apps have continued to appear on Meta platforms over the past few years. Researchers have found thousands that Meta failed to detect. Meta has also said it removed 344,000 ads on Facebook and Instagram promoting those services between November last year and January, and it has filed a lawsuit against at least one firm promoting nudify services.
Cindy Southworth, head of women’s safety at Meta, described sexual exploitation and the promotion of nudify apps as “horrific.” She said Meta is removing violating ads, launching new AI technology to detect and block new violating ads, and taking legal action against nudify app developers. She also said Meta shares information about removed apps and developers with other tech companies.
Southworth said the benign cover-photo tactic is an adversarial method used by advertisers to avoid detection. Meta is aware of the tactic, she said, and removes many violating ads despite those attempts to hide prohibited material. She added that no system is perfect and that Meta is investigating why these ads bypassed its systems.
Apple also removed Kromix
Apple’s role is also part of the story because Kromix had been hosted in the App Store before WIRED sought comment. An Apple spokesperson told WIRED that the company removed Kromix from the App Store.
Apple said the app appeared to have added prohibited content and features after initial review. The spokesperson said Apple has always strictly prohibited apps designed to generate, distribute, or consume pornography, and said nudification apps violate its App Review Guidelines.
The source article notes that Apple has struggled to keep similar tools off the App Store. Earlier this summer, San Francisco’s city attorney sent Apple a cease-and-desist letter demanding the removal of 13 specific apps and demanding that it stop “aiding and abetting” the sale of explicit deepfake images.
The Kromix case follows another recent WIRED report involving more than 50 ads on Meta’s platforms that contained AI child sexual abuse material and images of minors beside sexually suggestive statements. Those ads ran between November 2025 and August and sometimes reached thousands of people, including users in the US, UK, and at least a dozen European countries.
After WIRED contacted Meta about those ads last week, the company removed them from its ad library and replaced them with notices saying they had been taken down for violating policies on child sexual abuse and exploitation material, adult sexual solicitation, and nudity.
Together, the incidents show a difficult enforcement problem for major platforms. Meta and Apple both have rules that appear to cover this behavior. The harder question is whether those rules can be applied quickly enough when AI tools, ad accounts, app features, and evasion tactics keep changing.