How did Meta allow ads for deepfake porn apps?
Meta recently ran ads for an AI-powered app that generated deepfake pornographic videos of female politicians, despite the company's policies against such content. The ads were targeted at male users and featured a voice-over promoting the app as having "no restrictions." The incident raises questions about Meta's ability to regulate harmful content on its platforms.


Meta platforms recently ran ads for Kromix, an AI-powered app that let users create deepfake videos of female politicians and other women - and it's raising some serious red flags. These ads, targeted at male users, featured a voice-over touting the app's "no restrictions" and a tagline that promised to "Unleash Next-gen AI Magic." But here's the thing: one of these ads actually included a pornographic video with a deepfake that closely resembled a US politician. It's no wonder this incident has sparked concerns about Meta's ability to regulate harmful content on its platforms, especially when it comes to nonconsensual intimate imagery.
The Kromix app, which was hosted by Apple's App Store, was pretty brazen - it invited paid users to upload photos so that it could use them in some pretty disturbing scenarios, including "bedroom rape" and "Disney love." And if that's not enough, the app also featured AI-generated pornographic videos of women wearing Spider-Man costumes. Apple's had a tough time keeping similar tools off the App Store, and the company's faced criticism for its handling of nonconsensual intimate imagery. The ads were eventually removed after WIRED asked for comment, but the incident has raised some big questions about how Meta will deal with similar content in the future.
The use of deepfake technology to target female politicians and other women is a growing concern - and it's not hard to see why. For years, politicians and elected officials (who are overwhelmingly women) have been targeted with nonconsensual deepfake images and videos. The thing is, the production and distribution of AI-generated content have become so democratized that anyone with an internet connection can create and share harmful content. Meta's ad review system, which relies mostly on automated tools, failed to catch the ads for the Kromix app - and that's a pretty big problem. It highlights the need for more effective regulation of harmful content on social media platforms. So what can you do? Well, for starters, you can report suspicious ads and advocate for stronger policies against nonconsensual intimate imagery.
Source: Ars Technica
Related posts

Reson8 Releases Rumble-2-Flash 0.5B Speech Model for 40 Languages
Reson8 released Rumble-2-Flash, a 0.5-billion-parameter speech recognition model covering 40+ languages. Launched November 18, 2026 on Hugging Face under Apache 2.0, it runs in standard Transformers and vLLM without custom code. Distilled from the larger Rumble-2 via the GOLD on-policy method, it scores 6.96% error on the multilingual Open ASR Leaderboard, beating Meta's 7B omniASR-LLM and Whisper large-v3, and processes an hour of audio in about half a second on a single GPU.

Osmo Claims Its AI Model Matches Human Smell Ability
Osmo, a Google Brain spinout, claims its AI model achieves "human-level olfactory intelligence," detecting and discriminating smells like an average person and creating fragrances comparable to a junior perfumer. Founder Alex Wiltschko says the tech, trained on over ten million human scent labels, is already used by major fragrance firms and identified repellents stronger than DEET. Critics argue marketing claims lack documentation and the system struggles with contextual smell interpr

Google Launches Free Tasks AI Agent Replacing Spark
Google has rebranded its Spark AI agent as Tasks and made it free for users in the US, Canada, Japan, Korea, Australia, and India. Powered by Gemini 4 Hydrogen, the agent handles planning chores like meal plans and calendar organization. US users gain additional capabilities, including calling businesses, making purchases via Google Pay, and delegating tasks via email or SMS. Each session runs in an isolated virtual machine, pausing for confirmation before sensitive actions.