Meta Took Days to Remove Ads That Used AI to Create Fake Sexual Images of Real Children
An investigation found 332 ads on Facebook and Instagram this year promoting apps that digitally strip or sexually alter photos of real minors, including images taken from children's public social media profiles.
Key points
- The Tech Transparency Project found 332 ads containing AI-generated child sexual abuse material on Meta's platforms in 2025.
- The "vast majority" of the ads promoted AI apps made in China, including so-called "nudify" apps.
- Meta took several days to remove the ads after they were flagged, according to the investigation.
- Investigators matched ads to photos of real, named children, including a European royal family member and a 14-year-old Instagram influencer.
- The US Justice Department has stated that AI-generated child sexual abuse material carries the same legal weight as real images.
Researchers caught hundreds of ads running on Facebook and Instagram that promoted apps designed to generate fake sexual images of children using AI. The Tech Transparency Project, a nonprofit watchdog group that investigates how major tech platforms are used, published its findings on Tuesday. Meta, the company that owns both platforms, took several days to remove the ads after they were identified.
What exactly were these ads showing?
The ads promoted "nudify" apps, tools that use AI to digitally remove clothing from photos or animate still images into explicit videos. They are not niche products. They are marketed openly, and the investigation found 332 such ads slipping through Meta's content filters this year.
The Tech Transparency Project matched several ads directly to photos of real children that were publicly available online. One ad took a press photograph of a young member of a European royal family and animated it into a video depicting her in a graphic sex act. A second ad took a photo of a 14-year-old girl who posts about her life on Instagram and animated it into a video showing her performing oral sex. A third ad used stock photos of a younger girl in a pink athletic outfit and morphed them into a video showing an adult male assaulting her.
These are not abstract harms. These are specific children whose publicly shared images were fed into AI tools and turned into abuse material.
The "vast majority" of the ads promoted apps built by developers in China, first reported by Ars Technica.
Is this illegal?
Yes, under US federal law. The Justice Department has clarified that AI-generated child sexual abuse material, images or videos that were never real photographs but were created or altered by software, is treated the same as material that documents real abuse. That means distributing or possessing it can carry serious criminal penalties.
Meta's own policies also prohibit this content. The problem, according to the investigation, is that the company's automated systems failed to catch 332 ads before they ran.
What does this mean for parents?
Any photo a child posts publicly can potentially be fed into one of these apps. A school sports photo, a birthday picture, an Instagram post: all of it is accessible to anyone who wants to misuse it. That is not a reason to live in fear, but it is a concrete reason to check the privacy settings on your children's social media accounts and consider making their profiles private rather than public.
Meta has not yet said how it plans to prevent similar ads from running in the future. The Tech Transparency Project is calling on the platform to do more than remove individual ads after the fact, arguing that the advertising systems themselves are approving content that should never reach the review stage at all.
