Meta’s advertising controls are now under fire again after ads for an artificial intelligence app that can generate fake nude images of real people were reported on Facebook and Instagram. Kromix, a purported “AI image styler” app, but alleged to be able to generate sexually explicit images of people without their consent, has faced criticism.
The case has raised questions about how Meta’s systems detect and block harmful advertising, particularly as AI-powered image-generation tools become more sophisticated. The controversy also comes at a time when Meta CEO Mark Zuckerberg is already defending itself in a major US legal battle on charges that Facebook and Instagram deliberately kept children engaged for longer periods in order to make money. Meta has denied those claims.
Kromix advertisements were seen on Meta-owned platforms, even though the company bans sexual and exploitative content. The ads said it had “no restrictions” and marketed it as “the AI that men actually use.” Such promotional materials have led to renewed doubts about whether Meta's advertising review systems are sophisticated enough to find applications that make it possible for users to use non-consensual sexual images.
One advertisement showed a fake explicit image of a woman resembling a prominent US female politician. The promotion first showed the woman standing in front of US flags and then sexually explicit imagery featuring a person with the same facial appearance. This is one of the most disturbing cases of generative AI: using a real person's identity or likeness to produce sexual material without permission.
The bigger concern goes beyond the existence of an AI application. Critics will be particularly worried that the promotional material is reaching out to users through Meta's advertising system. Meta’s business model is so much about advertising and, with the help of automated systems and other means of review, it has been able to screen advertisements on the website and on its platforms.
The Kromix controversy therefore raises an important question: if Meta already has rules against sexual and exploitative content, why were advertisements promoting potentially abusive AI capabilities able to appear in the first place?
WIRED reported that Meta removed some of the advertisements after the publication contacted the company about them. The response has led to questions about whether or not the company's systems could identify the problematic advertisements independently or whether it was only after the issue was brought to its attention.
The incident is another example of the growing challenge of generative AI. Because of this, tools that can make highly realistic images and videos are becoming more widely available, and platforms are still struggling to distinguish legitimate image-editing applications from services that enable harassment, impersonation, or non-consensual sexual images.
Apple has also taken action against Kromix. Apple removed the app from the App Store after it was contacted about it, 9to5Mac reported. The company said the app had added new features that were not allowed after its original review. The removal shows how even after it’s made it through an initial platform review, an application’s capabilities can change, and it presents a new challenge in the development of apps and technology companies.
Meta is already under pressure in terms of platform safety and this is a particularly big deal. Facebook and Instagram have been accused of spreading harmful content, and the rapid growth of AI-generated media makes it even more difficult to moderate that.
Meta’s Kromix dispute could therefore serve as an additional test of whether its policies are being effectively implemented. Rules against harmful or exploitative content are only one part of the problem. The other is to make sure that automated advertising systems, human reviewers, and enforcement mechanisms are able to spot harmful content before it gets to the users.
The episode also raises broader questions about accountability in technology. As AI applications become more powerful in terms of image and video generation and as app stores, social media companies, and AI developers are increasingly being pushed to protect them from being misused, so does it matter for the rest of the technology sector to be accountable?
In the end, the Kromix case is a warning about the dangers of AI-generated images. A convincing fake image can have far-reaching reputational and emotional consequences when someone's face is put in sexually explicit material without consent. Preventing such abuse will be more than simply removing an app or advertisement after the abuse has been reported.
Meta then may have a bigger problem: How can Meta explain how ads for a tool that is alleged to promote “unlimited image manipulation” passed through its advertising ecosystem? What could be the answer is very important for the company’s platform safety statements in light of the continued debate among legislators, regulators, and the public about the balance technology companies must strike between engagement, advertising revenue, and user protection.
