Two review gates, one nudify app: Meta sold the ads, Apple carried the product

Meta's platforms ran ads for Kromix, an AI porn generation tool whose creative seemingly encouraged users to make deepfaked videos resembling female US politicians, despite Meta's policy barring ads that contain sexual imagery ^[1][2][17]. Apple's App Store hosted the app itself until WIRED asked about it, so the same product cleared two separate pre-publication reviews and was withdrawn at both gates only after a reporter made contact ^[5][9][1].

The creative was not subtle. A voice-over in one video ad promoted the tool as having "no restrictions," said "all the characters are real people," and called it "the AI that men actually use" ^[3]. In another, a woman closely resembling a prominent female US politician stands in front of a US flag and the flag of the president under the caption "What if she moved?", winks, and the ad cuts to a woman with the same face in a pornographic video ^[4]. The app, marketed as an "AI image styler," invited paying users to upload photos for use in scenarios including "bedroom rape" and "Disney love," and carried AI-generated pornographic videos of women in Spider-Man costumes ^[2][5][6]. A contact listed in the app did not respond to requests for comment ^[6].

Meta's own documentation says every ad is reviewed against its policies before it runs, and that "our ad review system relies primarily on automated tools" ^[15]. That system approved 32 ads from an account whose page had been created on August 3, shortly before the campaign began, and had no followers, two signals WIRED notes might have flagged it as suspicious ^[10][16][3]. The ads were targeted exclusively to male users under the tagline "Unleash Next-gen AI Magic," ran for between five and 46 hours, and mostly drew 10 or fewer impressions each ^[9][10]. On those figures the measured reach was small, in the low hundreds of impressions for the bulk of the campaign ^[2]. The defect is the gate, not the traffic: Meta's adult sexual exploitation rules already prohibit nonconsensual intimate imagery and the promotion of "nudify" or "AI kissing" apps ^[17], and researchers have found thousands of nudify ads the company failed to detect over the past few years ^[18].

Katie Paul, director of the Tech Transparency Project, which found the ads, said every nudify ad the group has seen so far exclusively sexualizes women, and questioned how Meta will handle the content when it is actively running such ads about female politicians ^[11]. Apple's record at its own gate is comparable: San Francisco's city attorney sent the company a cease-and-desist letter earlier this summer demanding the removal of 13 specific apps and that it stop "aiding and abetting" the sale of explicit deepfake images ^[7][8].

The supply side is well documented. Officials, overwhelmingly women, have been deepfake targets for years ^[12], and in January, as Grok was used to produce around 3 million sexualized images of women and girls, officials in Sweden, England and Northern Ireland were targeted, often starting from official images ^[13]. Benjamin Shultz of the American Sunlight Project says official portraits and press conference footage are now raw material for low- and no-code tools ^[14].

Watch whether Meta's automated review starts weighting cheap signals such as page age and zero followers before approving paid sexual-content ads ^[15][16], whether Apple's app review catches upload-your-photo apps at submission rather than after press calls ^[5][7], and what follows the San Francisco demand beyond the 13 named apps ^[8].