Apple has removed the Kromix app from the App Store after determining that it appeared to offer prohibited content and features. An Apple spokesperson told WIRED that the company maintains a strict ban on apps designed to create, distribute, or access pornography.
“Nudity” apps violate Apple’s App Review Guidelines, the spokesperson said, adding that the company actively works to identify and remove both offending apps and developer accounts that attempt to bypass its policies.
The removal follows a recent investigation by researchers at the Tech Transparency Project, as reported by WIRED. The group found that dozens of ads on Meta’s platforms promoted sexually suggestive services, including tools capable of generating abusive AI content involving minors. The ads reportedly ran between November and August 2025 and could have reached users in the United States, the United Kingdom, and at least a dozen European countries.
After WIRED contacted Meta, the company removed the ads from its advertising library and labeled them as violations of its policies covering child sexual abuse and exploitation, adult sexual solicitation, and nudity. Some of the ads appeared after Meta announced new artificial intelligence systems intended to improve the detection and blocking of policy-violating ads during upload.
Several ads were connected to an app called MaskAI, which was available in the App Store at the time. Like Kromix, MaskAI appeared to offer image-upload and face-swapping tools that could be used to create sexualized content. Apple removed the app after WIRED requested comment. An email address associated with the app did not respond to questions.
Paul said Meta is relying increasingly on artificial intelligence to screen and review advertisements, but argued that the company’s safeguards have failed to prevent harmful content from appearing. “Advertising is Meta’s main product, but there doesn’t seem to be any quality control,” she said. Meta has repeatedly highlighted its AI-based ad-review systems, although the company acknowledges that automated moderation is not perfect.
Additional reporting by Tim Marchman.
Source: arstechnica.com