Meta's platforms recently ran advertisements for an AI porn-generation tool that appeared to encourage users to create deepfake videos resembling female US politicians. This happened despite the company's policies that prohibit ads containing sexual material.
The tool, called Kromix, markets itself as an "AI image styler." According to WIRED, a voice-over in one of the video ads promotes the tool as having "no restrictions," saying "all the characters are real people" and describing it as "the AI that men actually use."
Deepfake Ads Targeting Female Politicians
One of the advertisements featured a pornographic video with a deepfake closely resembling a prominent US politician. In another ad, a woman resembling a female US politician appeared in front of the US flag and the flag of the president of the US.
This incident is the latest in a series of failures by Meta to keep advertisements for tools that produce nonconsensual intimate imagery off its platforms. The ads slipped through despite Meta's stated policies against sexual content in advertising.
"All the characters are real people." — Kromix ad voice-over, as reported by WIRED
Apple Removes the App
Following the reports, Apple removed the Kromix app from its store. According to coverage of the WIRED report, one advertisement featured a pornographic video with a deepfake closely resembling a prominent US politician, and Apple subsequently took action to remove the app.
The situation highlights a broader problem with AI-generated nonconsensual intimate imagery. As reported by AI Weekly, Meta ran ads on Facebook and Instagram for the app that promised to nudify female politicians, using a deepfake of a real person in the promotional material.
Meta's Ad Review System Under Scrutiny
The fact that these ads ran at all raises serious questions about Meta's ad review process. The company has policies in place that should have blocked this content, yet the ads still appeared to users.
This is not an isolated incident. The pattern suggests that Meta's automated and human review systems are not catching these violations consistently, allowing harmful content to reach audiences before being flagged.
- Kromix marketed itself as an "AI image styler" while promoting nonconsensual intimate imagery
- The ads used real people's likenesses without consent, including female politicians
- Meta's policies prohibit sexual content in ads, but these slipped through
- Apple removed the app after the reports surfaced
Our Take: Meta Must Do Better on AI Deepfake Ads
To put it plainly, this is a serious failure on Meta's part. The company has repeatedly said it takes nonconsensual intimate imagery seriously, but its advertising platform continues to let these tools through.
The use of real politicians' likenesses in these ads makes the violation even more egregious. These are public figures being sexualized without their consent, and Meta's systems did nothing to stop it until the ads were publicly exposed.
In our view, Meta needs to do more than just remove individual ads after they are reported. The company must implement stronger pre-screening measures that catch these violations before they reach users. Relying on post-hoc removal is not working, and the repeated nature of these failures shows that the current approach is insufficient.
For readers, this story is a reminder that AI tools can be misused in ways that cause real harm. The technology itself is not the problem — the problem is platforms that fail to enforce their own rules and protect people from having their images used without consent.
