Meta’s AI‑generated child abuse ads kept running in India despite a government ban
Meta ran hundreds of paid adverts that featured AI‑generated child sexual abuse imagery, including dozens in India, over the past ten months. The ads were discovered by the US‑based Tech Transparency Project (TTP) and confirm earlier warnings from a BBC investigation. The story matters because it exposes a gap between platform policies, legal obligations and the reality of automated ad screening.
The scale of the problem
According to BBC News, TTP identified 332 adverts on Facebook and Instagram that displayed child sexual abuse material (CSAM). The majority – 274 ads – appeared in August alone. Of the total, 84 ads ran in India, and 78 of those were posted after the Indian government ordered Meta to purge CSAM‑related content. The ads usually began with a real‑child photo that was then altered by artificial‑intelligence tools to depict graphic sexual acts. Most of the adverts linked to AI image‑generation apps, many of which are developed in China.
| Region | Total ads found | Ads in India | Ads posted after India’s ban |
|---|---|---|---|
| United States | 92 | 0 | 0 |
| United Kingdom | 41 | 0 | 0 |
| European Union | 45 | 0 | 0 |
| Australia | 20 | 0 | 0 |
| India | 84 | 84 | 78 |
| Other | 50 | 0 | 0 |
The table shows that India accounts for roughly a quarter of all CSAM ads, and the bulk of them appeared despite a clear regulatory directive.
How Meta’s ad review is supposed to work
Meta’s policy states that ads must not contain content that sexually exploits or endangers children. In practice, each ad first passes through an automated moderation system. The system scans images, video, text and audio, checks the destination URL, and evaluates the targeting parameters. If the AI flags a violation, a human reviewer can intervene. Advertisers can also be blocked if they repeatedly breach the rules. However, the system only evaluates ads while they are live; once an ad is taken down, reporting it retroactively is not possible.
India’s legal backdrop and recent actions
In India, the distribution of CSAM is a criminal offence under the Information Technology Act. After the BBC’s earlier investigation, the Indian government ordered Meta to remove all CSAM‑related ads and content on Instagram. The National Commission for Protection of Child Rights and the National Human Rights Commission have opened inquiries. Child‑rights activist Bhuwan Ribhu has petitioned the Supreme Court for mandatory reporting requirements, arguing that Meta is not complying with Indian law that obliges platforms to forward illegal content to local law enforcement.
The enforcement gap: why ads slip through
The hidden trade‑off is the reliance on post‑live detection combined with a global reporting pipeline that routes CSAM alerts to the U.S.‑based National Center for Missing & Exploited Children (NCMEC) rather than to Indian authorities. This means that even after a ban, ads can be published, observed for a short window, and then removed before any Indian law‑enforcement agency is notified. In TTP’s experience, Meta removed the ads only after they were reported, and sometimes responded that the ads “didn’t go against our Advertising Standards.” The response suggests a mismatch between the platform’s internal standards and the legal definition of CSAM in India. Moreover, many of the offending accounts are run by reseller partners in China, adding a layer of jurisdictional complexity that hampers rapid takedown.
What to watch next and how to act today
- Regulatory pressure: Expect the Indian Supreme Court to rule on the petition for mandatory local reporting. A clear mandate could force Meta to build a direct reporting channel to Indian police, bypassing NCMEC.
- Platform changes: Meta may upgrade its pre‑live screening to include more robust AI models trained on CSAM indicators, especially for content that starts with a real child photo and is later manipulated.
- Reseller scrutiny: Authorities could target the Chinese ad‑reseller network that funnels billions of dollars of spend onto Meta platforms. Cutting off these pipelines would reduce the volume of offending ads.
- Public vigilance: Users can still report ads through Meta’s built‑in tool while the ad is live. Reporting early increases the chance of swift removal.
- Immediate step: If you encounter a suspicious ad on Facebook or Instagram, click the “Report ad” link, select “It’s illegal” → “Child sexual exploitation,” and note the ad’s URL. Share the details with local child‑protection NGOs or the Cyber Crime cell to create a documented trail.
