New Delhi: Meta continued to carry paid advertisements featuring artificial intelligence (AI)-generated child sexual abuse material on Facebook and Instagram, including in India, according to a new report by the Washington-based Tech Transparency Project (TTP).
Indian agencies, including the National Commission for Protection of Child Rights and the National Human Rights Commission, have since taken up the issue. Among the children whose photographs were identified was a member of a European royal family. The report found 332 such advertisements on Meta’s platforms over the past 10 months. A large majority (274) appeared only in August. India accounted for 84 of the ads, with 78 appearing after the government had ordered the company to remove content and advertisements promoting child sexual abuse material from Instagram. The group said it reported all 332 advertisements to the National Center for Missing and Exploited Children (NCMEC), an international reporting system for child sexual exploitation online.
The TTP said the advertisements were identified across the United States, the United Kingdom, the European Union, Australia and India. The TTP report said several advertisements used photographs of real children and then changed those images using AI.
The findings come two months after the government order, which followed an earlier media investigation that found paid Instagram advertisements promoting child sexual abuse material in India.
Meta added that several of the advertisements reported by the TTP had already been removed. “We don’t tolerate nudify apps or any kind of child exploitation, whether real or AI-generated. “We have built strong defences. The TTP also tested Meta’s own system for reporting advertisements.
Because meta only provides the reporting option while an advertisement is live, the group said it tried to report all 84 India-based advertisements but could submit reports on 55. The 78 India-based advertisements that appeared after the government order have become a major part of the latest findings. The company said most had received fewer than 200 impressions and that the total amount spent on them was below $5,000. Of those 55 advertisements, the TTP said the company told researchers that 43 did not violate its advertising standards. The company said 11 had been removed, while there was no response in one case.
The report said around a quarter of all the advertisements it identified were shown in the country. Meta said it does not allow nudify applications or any form of child exploitation, whether the material involves real images or AI-generated content. These criminals constantly shift tactics to evade detection, which is why we keep strengthening our detection and enforcement,” the company said. It also said it has systems built to block advertisements that violate its rules and continued monitoring after advertisements were published. These include systems that block violating ads and continuously monitor them in case something initially gets past us,” the company said. The TTP said that by the time it sought a response from the social media company, all the advertisements it had identified had been removed.
Problem of users moving between platforms
Its founder Bhuwan Ribhu said Indian law requires such material to be reported to investigative agencies. The group also found that many of the advertisements directed users towards AI applications capable of generating manipulated images and videos. Meta’s advertising system reviews several parts of an advertisement before it is shown to users. These include images, video, text, audio, the audience being targeted and the link to which users are directed. The group has called for platforms to be required to identify accounts involved in spreading such material and report them to law enforcement agencies. The National Human Rights Commission has also sought reports from the Ministry of Electronics and Information Technology, the Ministry of Information and Broadcasting and the Delhi Police over allegations concerning paid advertisements on Instagram. “A major challenge with CSAM is the ease with which users can move from one platform to another. She added, “We hope the government will discuss this issue with social media intermediaries and give them appropriate directions. The findings also show a network of AI applications promoted through Meta’s advertising system. TTP’s latest findings show that advertisements promoting such applications continued to reach users in India even after government action and direct notification to the company about the material.
The TTP shared details of 17 advertisements with media organisations, including 11 that it said were running in India on September 1 and 2. Just Rights for Children, a network of more than 250 organisations working on violence against children in India, has sought directions from the Supreme Court on how social media companies should deal with child sexual abuse material.
The TTP said the advertisements began with an ordinary photograph of a child before using AI to create sexually abusive imagery. The organisation said it found new advertisements after informing Meta about its research. It also said it has systems built to block advertisements that violate its rules and continued monitoring after advertisements were published. “Under our child protection laws, it is mandatory to report child sexual abuse material to Indian investigating agencies, but by not doing so, social media companies are violating Indian laws,” he said. Shikha Goel, director of the Telangana Cyber Security Bureau, said one of the major challenges in dealing with such material is the ease with which users can move from one platform to another. This makes it difficult to catch them,” she said. The TTP said many of the advertisements directed users to applications developed by Chinese companies. The group said several advertisements had been run through China-based advertising partners, known as resellers, which place advertising on Facebook and Instagram on behalf of other businesses. These advertising partners, the TTP said, it contacted did not respond to its questions. Meta has said it prohibits advertisements promoting applications that generate fake nude or sexually explicit images without consent.
