Apple has removed an App Store app called Kromix that let paying users drop real people’s photos into AI-generated pornographic videos, after WIRED reported that Meta had been running Facebook and Instagram ads for it. One of those ads showed a woman closely resembling a prominent female US politician standing in front of a US flag, then cut to a pornographic deepfake with the same face. Meta pulled the ads and Apple pulled the app, both after WIRED asked about them.
Kromix listed itself as an “AI image styler.” The ads were less subtle: a voice-over promised “no restrictions,” said “all the characters are real people,” and called it “the AI that men actually use.”
What Meta’s ad system missed
The ads were found by the Tech Transparency Project, a watchdog group. According to Meta’s own ad library, they were targeted only at male users. Meta says the account behind them ran 32 ads, each live for between five and 46 hours, and most were seen ten times or fewer. The Facebook page that ran them was created on August 3 and had no followers, two signals that should have looked suspicious to Meta.
Most of the video ads opened with a harmless cover image, like a beach or a waterfall, before switching to the explicit content. Meta says it knows this trick and blocks many ads that use it, but “no system is perfect,” and it is still investigating how these got through. Meta’s rules already ban sexual content in ads and specifically prohibit promoting “nudify” apps. By its own count, Meta removed more than 340,000 nudify ads between November and January and has sued at least one nudify developer, yet researchers keep finding new ones.
Apple says the app changed after review
Apple told WIRED that Kromix appeared to have added the prohibited features after it passed initial App Review. Its statement repeats the usual line: Apple has “always strictly prohibited apps designed to generate, distribute, or consume pornography,” nudification apps break the App Review Guidelines, and it goes after both the apps and the developer accounts behind them.
That bait-and-switch pattern is the hard part for Apple. An app can look like a generic photo tool on submission day and turn into something very different after an update or a server-side change, and App Review does not catch every version. Apple ran into a similar problem with Grok earlier this year. After users generated millions of sexualized images and senators pushed Apple and Google to drop X and Grok, Apple warned xAI to fix the chatbot’s deepfake output or risk removal rather than pulling it outright.
Kromix was not the only removal. Apple also pulled MaskAI, a face-swap app that let users put people into sexual scenes. Some of the Meta ads pointing to MaskAI were among the more than 50 ads containing AI-generated child sexual abuse imagery that the same researchers surfaced earlier this month.
This lands a month after San Francisco’s warning
In July, San Francisco City Attorney David Chiu sent Apple and Google cease-and-desist letters over 13 nudify and face-swap apps, eight of them on the App Store, and gave both companies 28 days to respond. Apple said at the time it had removed three of the named apps and was shutting down their developer accounts. Kromix and MaskAI show that new ones keep arriving through the front door, and that Meta’s ad network is helping them find customers.
If you run into an app or ad like this, you can report it from the App Store listing or from the ad’s menu on Facebook and Instagram. The bigger fix, catching these before they go live, is on Apple and Meta.
