Meta is facing fresh scrutiny over its advertising enforcement after researchers discovered thousands of AI-powered 'nudify' ads running across Facebook and Instagram. The ads, which promote services that use artificial intelligence to digitally remove clothing from images of women, appear to have bypassed Meta's policies against sexually exploitative content.
The findings come from the Tech Transparency Project (TTP), an advocacy group that has repeatedly investigated whether major digital advertising platforms are effectively policing harmful content. According to TTP, a Chinese advertising partner named GatherOne placed more than 7,500 ads promoting such services on Meta's platforms. One of the promoted apps had previously been linked to child sexual abuse material, according to the researchers, yet users could still reach it through app stores or Meta's ad links.
The report raises significant questions about Meta's ability to enforce its content policies, especially as generative AI tools make it easier to create fake, sexually explicit images. The company has repeatedly said it uses a combination of automated systems and human moderators to remove content that violates its rules on sexual exploitation and harassment. However, TTP said that while Meta removed a large number of the ads, it appeared to have made a deliberate choice to continue doing business with GatherOne.
'This may reflect the company's desire to keep tapping into the massive Chinese advertising market, regardless of the impact on users,' TTP wrote in a blog post. Although exact figures are difficult to determine, TTP estimated that GatherOne could be spending more than $70 million per year on Meta ads, making it a substantial source of advertising revenue. The scale of the spending may help explain why the company has not cut ties with the partner.
What are nudify apps?
Nudify applications use machine learning models to generate realistic nude images from ordinary photos. They are often marketed to men as a way to 'see anyone naked' and are overwhelmingly used against women and girls. Critics have described these tools as a form of image-based sexual abuse because victims rarely give permission and the fake images can be shared, used for harassment, or sold. Some services are even promoted as a way to create content for adult websites, while others are operated as subscription-based mobile apps.
Such apps carry significant legal and ethical risks. In the European Union, the proposed AI Act includes requirements for transparency and safeguards around deepfakes. Several countries have passed or introduced laws that specifically criminalize the creation and distribution of non-consensual intimate images, including those generated by deep learning. Yet many of these services are hosted in jurisdictions with weaker enforcement, and they often rely on major platforms to attract users.
The TTP investigation illustrates how AI-generated abuse has moved from a fringe problem to a mainstream challenge for social media companies. Advertisers who operate at scale can run campaigns across multiple accounts and constantly modify their ads to avoid detection. Even when a particular ad is removed, the same advertiser can create a new one with slightly different wording or creative elements.
Meta's response and business ties
In a statement to Bloomberg News, Meta said it takes 'aggressive steps' against nudify apps. The company added that advertisers who violate its policies may face ad removals, financial penalties, and the suspension or termination of their advertising accounts. Meta also noted that it continues to invest in automated detection systems and human review teams to identify harmful content.
However, the TTP report argues that Meta's actions do not go far enough. GatherOne told TTP that it had stopped creating new advertising accounts for services related to nudify and face-swap tools, but it did not say it would stop existing campaigns or refund advertisers. TTP also noted that GatherOne appeared to remain an active advertising partner, suggesting that Meta was willing to tolerate some level of policy violations in exchange for revenue.
This is not the first time investigative groups have found sexual or illegal ads slipping through automated systems. TTP has previously documented similar issues on Google and Apple platforms, where thousands of ads remained live after user reports. The persistence of such ads points to structural problems in how ad tech companies screen campaigns. Automated filters are often based on text keywords, which can be bypassed with emojis, code words, or images. Human review is expensive and slow, especially when advertisers constantly generate new creatives.
Meta's AI troubles
Meta has also struggled with its own AI products. In July 2026, the company was forced to pull a feature from Muse Image, its generative AI image tool, just days after launch. The feature allowed users to reference public figures and place their faces on other bodies or into prompted scenarios. Actor unions, including the Screen Actors Guild of America, strongly criticized the feature because it was automatically enabled for public Instagram and Facebook accounts. Meta removed the feature shortly afterward, possibly fearing legal action.
This followed an earlier controversy involving AI chatbots. In 2025, Meta launched chatbots on its messaging platforms that allowed users to converse with AI characters resembling celebrities such as Taylor Swift, Anne Hathaway, and Scarlett Johansson. Users could configure these chatbots with flirtatious personalities and ask them to generate images of the celebrities in scantily clad outfits. Those chatbots were later disabled, but the incidents demonstrated a pattern of Meta shipping AI features without sufficient privacy or legality safeguards.
The company's attitude toward AI is shaped by intense competitive pressure. Meta and CEO Mark Zuckerberg have argued that AI is the next major computing platform and have invested heavily in data centers, chips, and research. Rivals such as Google, Microsoft, OpenAI, and TikTok parent ByteDance are also moving quickly to integrate AI into their products. In this race, Meta has sometimes appeared more willing to take risks than some of its peers.
Regulatory and legal consequences
The latest revelations may intensify regulatory scrutiny in several jurisdictions. Meta already faces lawsuits related to alleged addictive design practices targeting children. A lawsuit in New York is proceeding, while another case was dropped by the plaintiff shortly before trial. The European Union is reportedly considering a similar investigation, which could result in significant fines or mandatory changes to Meta's services.
Lawmakers in several countries have shown a growing appetite for restricting social media use by minors. Australia, Malaysia, and France have introduced social media bans, with France setting the threshold at under 15 and the other two at under 16. Turkey, Spain, and the United Kingdom have passed legislation, and several other countries are preparing similar measures. The EU may adopt a common social media ban for children, which would remove a significant number of young users from Instagram and Facebook.
While those debates focus largely on age-appropriate access, the nudify ads problem adds another layer of concern. Safety advocates argue that if Meta cannot prevent illegal ads from reaching users, regulators may consider stricter rules for advertising platforms. That could include independent audits, mandatory takedown deadlines, or liability for ad revenue earned from ads that violate the law.
What could happen next
Meta's response to the TTP report will be closely watched by advertisers, regulators, and civil society groups. The company has not announced any changes to its relationship with GatherOne. It also has not publicly commented on the specific ads identified by TTP beyond its general statement about enforcement.
For now, the burden remains on researchers and journalists to expose such failures. TTP's findings show that harmful AI-generated content is not only circulating on social media but is being actively monetized by advertising partners. The fact that one partner can spend tens of millions of dollars while running thousands of violating ads suggests that Meta's enforcement is not as robust as the company claims.
The broader technology industry faces a similar challenge. As generative AI tools become more powerful, the line between legitimate content and harmful deepfakes will continue to blur. Advertising networks are often the economic engine behind these services, giving them a unique responsibility to police the content they distribute. The coming months will likely bring more investigations, more lawsuits, and perhaps new rules that force platforms to take non-consensual intimate imagery more seriously.
Whether Meta changes its advertising enforcement remains to be seen. However, the TTP report is likely to intensify questions from regulators and safety advocates over how effectively the company detects AI-enabled abuse and whether repeat advertisers face meaningful consequences for violating its policies.
Source: TechRepublic News