AI nudify tools have made it easier to turn photos of real people into fake nude images. The problem is worse when social media platforms promote AI nudify apps that create sexual images without consent.
In August 2026, a WIRED investigation found that Meta had run ads for an AI app called Kromix. The app promoted features that could create pornographic deepfakes of real people. One ad showed a woman who looked like a well-known female US politician. The video then switched to an explicit clip featuring the same face.
This has raised concerns about Meta’s ad review process. They also questioned whether the company does enough to stop harmful AI tools from reaching users on Facebook and Instagram.
Meta prohibits sexual content in ads and restricts the promotion of apps that create nonconsensual intimate images. Yet, the Kromix ads still appeared on its platforms. This raises questions about how effectively Meta enforces its own advertising policies.
The issue goes beyond one app or ad campaign. Social media platforms must check what advertisers promote when they pay to reach users.
Meta Ran 32 Ads for an AI Nudify App
According to WIRED, Kromix presented itself as an AI image-editing app. However, its features went beyond basic photo editing.
The investigation found that paying users could upload photos and place people in different situations, including sexual scenes. One ad showed a woman who looked like a well-known female US politician. It then cut to a pornographic scene featuring a woman with the same face.
The ads also suggested that users could create these images with few restrictions on who they depicted. WIRED reported that the account behind the campaign ran 32 ads, according to Meta’s Ad Library. The ads stayed active for between five and 46 hours. Most received 10 or fewer impressions before they were removed.
The low number of impressions may have limited the campaign’s reach. But it does not address the main concern. Meta’s advertising system still allowed the ads to run, despite its rules against promoting this type of sexual content.
Meta removed the ads after WIRED contacted the company about them. This case raises questions about how Meta reviews advertisements before they appear on Facebook and Instagram. The number of people who see an ad is only part of the problem. Even ads with limited reach can expose gaps in the system used to review millions of advertisements.
Meta’s Ad Review System Failed to Stop the Campaign
Meta says it reviews ads before they appear on Facebook and Instagram. Its advertising policies and safety rules also restrict the promotion of nudify apps and content that creates intimate images without consent. However, the Kromix campaign raises questions about how well Meta enforces these rules.
According to WIRED, Meta relies mainly on automated tools to review ads. The page linked to the campaign was also created shortly before the ads appeared and had no followers. These details could have raised concerns. However, the report does not confirm whether Meta’s review system flagged them.
Automated tools help platforms review large numbers of ads quickly. But they can miss harmful content if advertisers find ways to hide what their products do. Meta needs safeguards that can identify these tactics before an ad reaches users.
WIRED also reported that some Kromix ads used harmless images of beaches, mountains, and waterfalls as cover images for videos containing pornographic material. These tactics may make it harder for automated systems to identify harmful ads during the initial review.
Meta’s head of women’s safety, Cindy Southworth, told WIRED that the company had removed more than 340,000 ads promoting nudify apps late last year. She also said Meta was improving its detection tools and taking legal action against app developers. The company acknowledged that some ads had bypassed its systems.
These steps are important, but they do not solve the problem on their own. Removing an ad after a journalist or researcher reports it is different from stopping it before it appears.
Meta needs to explain why the Kromix ads passed its review process. It should also show what changes it is making to prevent similar campaigns from appearing again. Stronger checks before publication could help stop ads promoting nonconsensual sexual images from reaching users in the first place.
ALSO READ: UK Bans Ads for AI Nudify Apps That Sexualize Real Women
Female Politicians Face a Different Risk From AI Deepfakes
AI-generated sexual images can harm anyone whose photos are used without consent. But for female politicians, the risks can extend beyond personal privacy. Such content can be used to attack their reputation and undermine their role in public life.
Female politicians regularly share photos, campaign videos, and interviews online. AI tools can misuse these public images to create fake sexual content without their permission.
The WIRED investigation showed how an image resembling a prominent US politician was used in an ad promoting pornographic content. The investigation did not establish that the politician herself appeared in or helped create the explicit material.
This highlights a key risk of deepfake pornography. Fake images and videos can make people appear connected to sexual content they never agreed to create. Even after removal, the material may already have been shared.
For women in politics, this abuse can damage their reputation, distract from their work, and discourage them from participating in public debate.
Meta should take these risks seriously. Its ad review system needs to do more to identify and block campaigns that promote nonconsensual sexual deepfakes, especially those that use recognisable faces to attract attention.
ALSO READ: Nudify Sites Target Nearly 150 European Politicians in Deepfake Porn Campaign
Meta Has Faced Repeated Problems With Nudify App Ads
The Kromix case is not the first time Meta has faced questions about ads promoting AI nudify apps. Previous investigations have also found advertisements for tools that can create sexual images of real people without their consent.
In June 2025, Meta announced legal action against the entity behind CrushAI. The company also said it was developing tools to detect ads promoting nudify services and sharing information about these apps with other technology companies.
However, similar advertisements continued to appear on Meta’s platforms. A CBS News investigation in June 2025 found hundreds of ads promoting AI tools that could create sexual deepfakes of real people.
In July 2026, the Los Angeles Times reported that the Tech Transparency Project had identified around 7,600 ads linked to a Meta advertising partner in China. The ads promoted apps that could create sexualised images and videos using women’s faces.
Meta has also reported removing large numbers of these ads. According to its first-half 2026 adversarial threat report, the company removed more than 344,000 ads promoting nudify apps from Facebook and Instagram between November 2025 and January 2026.
These figures show that Meta has taken action against such advertisements. But they also show the scale of the problem. Repeated investigations have continued to find similar campaigns despite the company’s enforcement efforts.
Meta needs to make it harder for advertisers to return with new accounts, different app names, or altered ads. It should also explain how many harmful advertisements its systems block before publication. Removal figures alone do not show how well its review process works or how many violating ads reach users first.
Apple Removed Kromix After WIRED’s Investigation
Meta was not the only company linked to the Kromix case. The app was also available on Apple’s App Store before WIRED contacted the company. Apple removed it after the publication asked questions about its features.
An Apple spokesperson told WIRED that Kromix appeared to have added prohibited content and features after its initial review. Apple said it bans apps designed to generate, distribute, or access pornography and removes apps that break its rules.
The case raises questions about how app stores and advertising platforms check AI tools. Apple decides which apps can remain on its App Store, while Meta reviews ads shown on Facebook and Instagram. Both companies have a role in stopping harmful AI apps from reaching users.
An app may initially appear to be a basic photo editor before adding features that enable sexual deepfakes. Its advertisements may also promote these features more openly than its app store listing.
Removing Kromix was necessary, but companies also need better ways to identify apps that change after approval. Closer cooperation between app stores, advertising platforms, and safety researchers could help prevent similar tools from spreading across services.
Meta Needs to Be More Accountable for the Ads It Runs
Meta has taken action against nudify apps through legal measures, improved detection tools, and ad removals. However, the Kromix case shows that the company needs stronger checks to stop harmful ads before they reach users.
Meta should strengthen advertiser checks and review apps more closely. New accounts, misleading thumbnails, suspicious ad language, and links to previously flagged services should trigger further review.
The company should also make it easier for researchers to report harmful ads and track what happens after a report. Its Ad Library provides useful information, but reports must lead to timely action.
Meta should publish clearer data on its ad moderation system. The number of removed ads does not show how many harmful ads escaped detection, how long they remained active, or whether the same advertisers returned with new campaigns.
Human reviewers should also play a clear role in cases involving nonconsensual sexual images. Automated tools can review large numbers of ads, but they need stronger support to catch advertisers who disguise harmful content.
These changes would help users, researchers, and regulators judge whether Meta’s advertising rules are working in practice.
