Australia’s eSafety Commissioner is investigating allegations that Meta continued to run hundreds of advertisements depicting child sexual abuse material (CSAM), including more than 120 ads that were allegedly targeted at Australian users.
The investigation follows a report from the Tech Transparency Project (TTP), which found 332 ads containing CSAM that had run across Facebook and Instagram this year.
According to the research, 120 of the ads ran in Australia, while 262 ran in the US and 84 ran in India. The majority of the ads identified by TTP ran in the European Union and the UK.
TTP said the ads featured children ranging from toddlers to teenagers, with many using AI to digitally manipulate photographs of children to make them appear to be participating in sexual acts.
The findings come just weeks after Meta removed around 50 ads identified by researchers as directly containing CSAM. At the time, Meta said most of those ads had been published before it introduced new AI technology designed to “better detect and block” harmful advertisements at upload.
However, TTP said its latest investigation found hundreds of additional ads that ran after those new safeguards were introduced.
In one particularly disturbing example, researchers identified an advertisement that used an officially released photograph of a young member of a European royal family. The image was digitally manipulated into a video depicting a graphic sex act and promoted a Chinese AI app.
TTP said it identified multiple photographs of real children that had been taken from social media posts, websites and stock imagery and subsequently used in the ads.
The research also found that 298 of the 332 ads directed users towards AI image or video-generation apps, with 253 pointing to apps developed by China-based companies. TTP said it verified that many of the apps promoted in the ads were capable of “nudifying” people or placing them into sexual videos.
Meta Australia declined to comment when reached by B&T. However, in a statement to tech publication WIRED, the global tech giant said “We don’t tolerate nudify apps or any kind of child exploitation, whether real or AI-generated.”
A spokesperson for Australia’s eSafety commissioner told Wired that it is “highly concerned” by the reports. They also said it is assessing Meta’s compliance with its rules. It says it has requested “information on this matter from the company at a senior level”.
TTP identified that several of the ads as having been placed by Meta’s Chinese advertising reseller partners.
Three resellers – GIMC, Meetsocial and BlueFocus – were linked to CSAM advertisements, according to the investigation.
GIMC, also known as Guangdong Advertising Group, is a state-controlled Chinese company and one of Meta’s authorised resellers in the country. TTP said Meta has previously recognised GIMC as an advertising “success story”. The investigation found that GIMC, Meetsocial and BlueFocus had each placed ads containing the same CSAM content.
TTP also claimed that Meta pays Chinese resellers a 10 percent commission on advertising purchased through their accounts, citing previous Reuters reporting that the company has whitelisted some reseller advertising and does not always act immediately when ads are flagged by its automated systems.
Beyond the content itself, TTP found apparent failures in Meta’s enforcement systems.
The researchers reported 129 active CSAM ads directly to Meta. And in 57 percent of cases, or 73 ads, Meta allegedly responded that the advertisements did not violate its Advertising Standards. In a further 56 cases, Meta said the ads had already been removed, although TTP said six remained visible in the company’s Ad Library.
TTP also found that some ads continued accumulating views while Meta reviewed reports. One advertisement featuring a real child had reached four users in the EU when it was reported, but accumulated a further 330 views before it was removed.
Meta also removed and reclassified a significant number of the advertisements following TTP’s investigation.
As of September 1, Meta had removed 183 of the 332 identified ads from its Ad Library. After TTP shared its findings with the company on September 2, Meta removed the remaining 149.
The researchers said Meta also changed the policy violation notices attached to 113 of the ads, which had previously been categorised under adult sexual content policies rather than child exploitation.
TTP said the misclassification raised questions about the accuracy of Meta’s own reporting around the volume of CSAM it detects and removes.
The group also pointed to potential transparency issues in the EU, where it said many of the advertisements failed to properly disclose who paid for or benefited from them.
The research comes as Meta continues to publicly position itself as a leader in online child safety.
In July, the company said it takes a “zero tolerance approach” to child exploitation and described those behind such content as “determined criminals”.
