Meta and its Chinese partner are at the center of a scandal over advertising AI locker rooms

The story around generative artificial intelligence received a new continuation, and the focus was not just another neural network, but the largest advertising platform in the world. Meta, which has been vocal about combating deepfakes and non-consensual intimate images for the past two years, has faced allegations that its Facebook and Instagram services spent months running ads for apps capable of creating fake nude photos. The scandal quickly expanded beyond a simple moderation error and raised questions about how effectively the tech giants police their own advertising ecosystems.

When it comes to deepfakes, people usually discuss already published images or videos, notes xrust. However, researchers from the American organization Tech Transparency Project (TTP) decided to look at the problem from the other side — who helps such services find new audiences. The results were unexpected.

According to TTP, over the past few months, about 7,600 advertisements , promoting mobile applications that use artificial intelligence technologies to create intimate images and videos. Some ads looked almost harmless, but were accompanied by characteristic slogans like “Create a video with any face” and photographs of young women in explicit poses. For the user, it looked like an advertisement for another AI service, although in fact it was about tools that allow you to create deepfakes of a sexual nature.

The researchers paid special attention not only to the advertisements themselves, but also to the person who posted them. A significant part of the advertising campaigns, as stated in the report, took place through the Chinese company GatherOne . She has a long history of working with international advertising platforms and helping Chinese businesses promote their products outside the country. On Meta's corporate website, GatherOne is among the eleven official partners accompanying Chinese clients' advertising campaigns in overseas markets.

This fact in itself does not mean a violation of the rules. Facebook and Instagram have been blocked in mainland China for years, but Chinese electronics manufacturers, game developers, online stores and other companies continue to buy ads on Meta services to promote products around the world. That is why large international businesses have formed around such intermediaries.

GatherOne emphasizes that they work not only with mobile application developers. The company's clients include representatives of the gaming industry, e-commerce, entertainment sector and financial market. In response to questions from reporters, the company said it has a zero-tolerance policy against fraud, child exploitation, sexually explicit material and any use of artificial intelligence to create intimate content without human consent.

Following questions surrounding the investigation, GatherOne said it was stopping the creation of new advertising accounts for services related to AI stripping and face replacement, and was also conducting an internal review of advertising policies taking into account the development of synthetic media technologies.

Meta also rejected accusations of inaction. The company recalled that the rules of the advertising platform prohibit the promotion of services that create intimate images without people’s consent, and advertisers are required to comply with uniform standards regardless of country of origin. According to Cindy Southworth, head of women's safety policy, the corporation consistently removes such content and blocks violators if they bypass existing restrictions.

However, the authors of the investigation believe that the practice is at odds with official statements. Tech Transparency Project director Katie Paul said the uncovered ad campaigns make it appear as though Meta, one of China's largest advertising partners, has long been treated more leniently by the platform than the company's public policy suggests. In her opinion, such cases undermine the credibility of the corporation’s statements about the fight against services that create intimate deepfakes.

One of the services studied by the researchers caused a particular resonance — the application BAfter for Android. It was advertised as just another image processing tool, but according to TTP, there was a section inside for sharing materials with face replacement technology. Users of the Google Play Store have left reviews claiming to have encountered highly questionable content, including possible content involving minors. The developer has not publicly commented on these accusations.

It is important that the application disappeared from Google Play on July 21 , even before the publication of the investigation. GatherOne later announced that it had permanently stopped advertising services to the organization associated with the service. Meta also confirmed the blocking of BAfter and a number of other applications, links to which are no longer allowed to be placed on the company’s advertising network.

Why the AI locker room story became a problem for the entire industry

The scandal surrounding the advertising of such applications turned out to be not only a question of the quality of moderation. He touched on a deeper problem: the speed of development of generative AI turned out to be faster than the ability of Internet platforms to create effective control mechanisms.

Just a few years ago, creating a realistic fake image required complex software and special skills. Today, such tools are available to almost any smartphone user. A few photos of a person from social networks are enough to get an image that can look convincing to others.

Teenagers have become a particularly vulnerable group. A study from George Mason University, which involved 557 teenagers from the United States , showed the scale of the problem: more than half of those surveyed said they had used artificial intelligence image generation tools to create pictures of themselves or other people. More than a third of participants reported that they themselves had been victims of such technologies.

According to digital forensics researchers, many teenagers already know where to look for such services: they are distributed not only through application stores, but also through websites, instant messengers and closed communities. Therefore, simply removing individual programs does not completely solve the problem.

The Chinese market remains an important source of income for Meta

The GatherOne story took on additional weight due to the financial significance of the Chinese direction for Meta.

Although Chinese residents do not have official access to Facebook and Instagram, Chinese companies are actively using the advertising tools of the American corporation to reach foreign buyers. According to Meta, its revenue in China for 2024 amounted to $18.4 billion — about 11% of the company's total global revenue . This figure is more than double the 2022 level.

This is why any issues related to Chinese advertising partners become not only a reputational, but also a financial problem for Meta.

Additional context emerged after the publication of a Reuters investigation in December. Then the agency reported that about 19% of Meta advertising revenue in China in 2024 may have been for ads related to prohibited categories such as fraud, illegal gambling, and pornographic content. Among the companies that were mentioned in connection with such advertising schemes was GatherOne.

Meta then stated that it pays great attention to the fight against rule-breaking advertising and is constantly improving verification systems.

Old questions about Meta's work with China

The new scandal also recalled previous disputes surrounding the relationship between Meta and the Chinese market.

The company's former public policy director, Sarah Wynne-Williams, who later became a whistleblower, testified before the US Congress last year. She alleged that Meta had offered censorship tools to China for years and helped the country advance in an artificial intelligence arms race against the United States. Wynne-Williams later detailed her experience at the company in the book Careless People, published in 2025. Meta previously rejected her statements, but the very topic of interaction between American technology giants and the Chinese market remains one of the most sensitive in the industry.

Now the problem of artificial intelligence and responsibility for the spread of dangerous tools has been added to the political and economic issues.

Regulators are trying to catch up with technology

Governments of different countries are gradually beginning to respond to the new reality.

Last year in the USA, a law was passed Take It Down , which introduces liability for the publication of intimate materials without a person’s consent and obliges Internet platforms to delete similar content. The Federal Trade Commission began enforcing the law in the spring.

However, experts believe that legislation faces the same problem as platforms: technologies change faster than new rules appear.

If earlier the main question was why the social network did not delete the already published material, now another question arises — why the advertising system even allowed the promotion of tools for creating such content.

AI is changing the rules of the game for advertising platforms

Meta's history shows that the problem of artificial intelligence has already gone far beyond conversations about beautiful images, text generation or new digital assistants.

For advertising platforms, the new reality means the need to check not only the content of the ad, but also the product it promotes. The app may look like a regular photo editor, but hide features that can create dangerous synthetic content.

According to researchers, the applications that TTP studied received a total of more than 2.5 million downloads worldwide . This means that we are not talking about isolated experiments by developers, but about services with a real audience.

Meta has already announced blocking some of these applications and strengthening control over advertising. GatherOne announced a review of its own policies. However, the very fact that thousands of ads are appearing shows that existing filters have not yet kept pace with the development of generative AI.

For technology companies, this is becoming a new challenge: the struggle is not only for the attention of users, but also for trust in the digital environment, where the border between a real image and an artificial fake is becoming less and less noticeable.

Based on materials from https://www.bloomberg.com

Xrust Meta and its Chinese partner are at the center of a scandal due to advertising AI locker rooms