NEW DELHI — When Meta’s own review team looked at one of the Instagram advertisements a BBC investigator had just flagged, it decided the ad did not violate the platform’s community standards. Twenty-four hours had passed since the report. India’s government took considerably less time to disagree.
The Ministry of Electronics and Information Technology issued a formal notice to Meta on Saturday evening, ordering the company to immediately disable advertisements and content the ministry said promoted or facilitated access to child sexual abuse material on Instagram. The notice, described by officials familiar with it, gives Meta seven days to explain in writing how the advertisements were approved in the first place, what corrective steps it has taken since, and what safeguards it intends to put in place to stop it from happening again.
The notice follows a BBC Eye investigation, published July 3, that set up a test account inside India and found the platform’s advertising system approving roughly 30 unique paid advertisements using explicit search terms, some directing users to third-party channels on the messaging app Telegram where illegal material was allegedly being sold for as little as 99 rupees, about a dollar. The investigators say the account had not searched for any of this material directly. It arrived through Instagram’s own recommendation and advertising systems, after the account had simply followed a handful of profiles posting suggestive content over the course of a week.
Meta’s first response to the specific ad the BBC flagged was to clear it. Only after BBC journalists contacted the company directly, rather than through the in-app reporting tool, did Meta say it had disabled several advertisements, suspended the accounts posting them, and blocked some of the external links. That sequence, an automated review clearing content that a second, human-driven inquiry could not, appears to be what India’s ministry is most focused on: the notice reportedly singles out the platform’s algorithmic amplification of the advertisements as a specific target of its inquiry, not just the individual ads themselves.
The seven-day deadline is not a formality. Under India’s Information Technology Act, platforms that fail to act against unlawful content after being notified risk losing the safe harbor protection under Section 79 that shields intermediaries from liability for what users post. The ministry has invoked that threat before, issuing similar notices to X, YouTube and Telegram over child sexual abuse material earlier this year and warning that continued noncompliance would trigger the same withdrawal of protection.
Meta has not said publicly whether it will meet the seven-day deadline, nor has it detailed what internal failure allowed advertisements using words as blunt as those the BBC described to clear its automated ad-review system in the first place. The company’s ad approval process is largely automated, screening submissions using pattern-matching systems that are not designed to catch every euphemism traffickers use, and Meta has acknowledged in past incidents that its systems miss content that evades specific banned keywords. A separate privacy lapse discovered in Meta’s smart glasses app in June followed a similar pattern: a capability quietly present in the product, surfaced only after outside scrutiny rather than caught by the company’s own review. What Meta has not explained, in this case or that one, is why an issue a human reviewer could evaluate correctly on a second look was cleared by the system meant to catch it the first time.
Child safety advocates in India have pushed for years for platforms to be held accountable not just for content users post but for what their own advertising systems approve and distribute, a distinction regulators internationally have struggled to enforce. India’s Protection of Children from Sexual Offences Act criminalizes both the material itself and facilitating access to it, giving the ministry a statutory basis for its notice beyond the IT Act’s intermediary rules. Whether that combination produces a different outcome than the notices sent to X, YouTube and Telegram is the open question hanging over Meta’s next seven days.
India has issued this kind of notice before without ultimately stripping any major platform of its safe harbor status. Whether this case ends differently, given the subject matter and the specificity of the ministry’s questions about algorithmic amplification, is not something either side has indicated. Meta’s written response is due within the week. What it says, and whether India considers it sufficient, has not been decided by anyone outside the ministry itself.
