AI
Meta Platform Hosted Ads for App That Generated Fake Nude Images of Female Politicians
The social media giant's ad network promoted an AI tool designed to create non-consensual intimate imagery despite explicit company policies against such material.
Ars Technica
Meta's advertising platform recently promoted an artificial intelligence application engineered to generate non-consensual sexual imagery, a discovery that highlights the company's ongoing struggle to prevent harmful tools from reaching users through paid promotion.
The Deceptive Ad Campaign#
The ads marketed an app called Kromix, which the company styled as an "AI image styler" but functioned as a tool for creating deepfaked pornographic content. Video advertisements viewed by external observers featured messaging like "no restrictions" and "the AI that men actually use," while showing imagery designed to resemble female U.S. political figures. One particularly revealing ad displayed a woman resembling a prominent female politician before the video transitioned to explicit material.
According to Meta's advertising records, the campaign exclusively targeted male users with promotional language about "Unleash Next-gen AI Magic." In total, the account responsible for the ads launched 32 different advertisements, with most generating minimal engagement—typically under 10 impressions each—before Meta removed them in response to inquiries.
The Broader App Ecosystem Problem#
The Kromix platform itself, which had been available through Apple's App Store, permitted users to upload photographs for use in scenarios explicitly labeled as "bedroom rape" alongside other similarly disturbing applications. The app included examples of AI-generated sexual videos featuring fictional characters. Despite clear platform policies prohibiting tools that generate non-consensual intimate content, both major app marketplaces have struggled to prevent such applications from remaining available.
Earlier this summer, San Francisco's city attorney issued a cease-and-desist letter demanding that Apple remove 13 specific applications of this nature and cease facilitating the distribution of such material.
Why This Matters#
This incident represents a systemic challenge in content moderation at scale. Meta has previously failed to prevent ads for tools generating non-consensual intimate imagery, including an earlier incident involving ads linked to AI-generated child sexual abuse material. The proliferation of these tools threatens not just privacy but also the integrity of public discourse, particularly for women in politics who face unique vulnerabilities to image-based harassment campaigns.
