Meta Ran Ads for Deepfake App Targeting Female Politicians

Meta recently approved and ran advertisements for an AI-powered application that explicitly marketed its ability to generate deepfake pornography resembling female U.S. politicians, according to findings from the Tech Transparency Project first reported by WIRED.

The application, called Kromix, described itself as an "AI image styler" but promoted features with "no restrictions" in video advertisements that ran on Meta's platforms. One ad featured imagery closely resembling a prominent female politician in front of U.S. and presidential flags, with text asking "What if she moved?" before transitioning to pornographic content. The voiceover claimed the tool featured "all the characters are real people" and marketed it as "the AI that men actually use."

How the Ads Evaded Detection

Meta's advertising system relies primarily on automated tools to review content against company policies before ads go live. The Kromix campaign ran 32 ads over periods ranging from five to 46 hours, with most receiving fewer than 10 impressions. The ads were exclusively targeted to male users.

The advertisements employed evasion tactics, using cover images of beaches, mountains, and waterfalls to disguise pornographic content from automated screening systems. The account running the campaign was created on August 3, shortly before the ads began, and had no followers—indicators that might have flagged suspicious activity.

Meta removed the ads after WIRED inquired about them. The company has explicit policies prohibiting sexual imagery in advertisements and banning promotion of "nudify" or "AI kissing" applications.

A Persistent Problem

This incident represents the latest in an ongoing pattern. Between November and January, Meta removed 344,000 ads promoting nudify services from Facebook and Instagram. The company has filed lawsuits against at least one firm promoting such services and shares information about removed apps with other technology companies.

"All of the ads that we've seen so far broadly on the nudify issue are exclusively sexualizing and nudifying and putting women into AI-generated porn," said Katie Paul, director of the Tech Transparency Project. "This raises a lot of questions about how Meta is going to deal with this kind of content, not just broadly but when it comes to female politicians."

Apple removed Kromix from the App Store after WIRED's inquiry, stating the app appeared to have added prohibited content after initial review. The app had invited paid users to upload photos for use in scenarios including "bedroom rape" and "Disney love." Apple has faced similar challenges; San Francisco's city attorney sent a cease-and-desist letter this summer demanding removal of 13 specific apps creating nonconsensual intimate imagery.

Why it matters

Female politicians and elected officials face disproportionate targeting through nonconsensual deepfake technology, with official portraits and press conference footage becoming raw material for abuse. As AI tools democratize the creation of synthetic media, the gap between platform policies and enforcement capabilities creates real harm for public figures—predominantly women—while undermining trust in visual media during election cycles. The reliance on automated moderation systems, even for Meta's "primary product" of advertising, demonstrates the limitations of AI-based content screening at scale.

Cindy Southworth, head of women's safety at Meta, acknowledged the challenge: "This is a highly adversarial space where bad actors constantly evolve their tactics, which is why we're continually strengthening our defenses." She noted the company has launched new AI technology to detect violating ads but conceded "no system is perfect."

Last week, WIRED reported that the Tech Transparency Project found more than 50 ads on Meta's platforms containing AI-generated child sexual abuse material, some of which ran after Meta claimed to have implemented improved detection technology.

These details were first reported by WIRED, with additional reporting by Tim Marchman.