For years, Meta has positioned artificial intelligence as the backbone of its content moderation systems. Billions of posts, reels, comments and advertisements pass through automated review every day, with the company repeatedly claiming that machine learning helps detect harmful content before users ever see it. Last week, that narrative faced one of its biggest tests yet. Following a BBC Eye investigation alleging that Instagram served paid advertisements promoting child sexual abuse material (CSAM) in India, the Ministry of Electronics and Information Technology (MeitY) has issued a notice directing Meta to immediately disable such advertisements and explain how they appeared on the platform in the first place.
The government has reportedly given Meta seven days to submit a detailed explanation covering the action taken, its moderation systems and the safeguards currently in place. But the larger question extends well beyond one investigation. If Instagram’s advertising system is among the world’s most sophisticated, how did paid advertisements allegedly promoting one of the internet’s most serious crimes make it through?
Also read: Meta Introduces Spin View, Multi-Cam Stories, And New Editing Tools For AI Glasses Users On Instagram
The Investigation Didn’t Begin With Illegal Content
One of the most striking findings from the BBC investigation is how ordinary the journey reportedly began. According to the report, the journalists did not search for illegal material. Instead, they began with what appeared to be a harmless video inside Instagram’s recommendation system. Over the following days, Instagram’s algorithm allegedly began recommending increasingly disturbing content.
Within a week, reporters claimed they were shown advertisements promoting child sexual abuse material.
Unlike user-generated posts that often appear organically, these were allegedly paid advertisements. That distinction fundamentally changes the conversation. Advertising passes through a different layer of moderation.
Before an advertisement reaches users, Meta reviews both creative assets and campaign details against its advertising policies.
Those systems are designed specifically to prevent illegal, harmful or policy-violating promotions from ever being approved. Yet the BBC alleges that some advertisements remained active even after they were reported. According to the investigation, Instagram responded that one reported advertisement did not violate its policies. If accurate, this points not simply to missed content but potentially to failures across multiple review stages.
Why Paid Advertisements Matter More Than Organic Posts
Content moderation is an impossible task at internet scale. Every major platform struggles to remove every harmful image, video or comment uploaded by billions of users. Advertisements are different.
Unlike ordinary posts, advertisements undergo approval before publication. They generate revenue for the platform.
They are reviewed using dedicated advertising policies alongside automated systems and human reviewers. This is why allegations involving advertisements often attract greater regulatory scrutiny than user-generated content. The question shifts from “Why didn’t Meta remove it?” to “Why was it approved at all?” That distinction explains why India’s response has been swift.
Meta’s Defence: A Constant Battle
Meta has strongly rejected suggestions that it knowingly allows such material on its platforms. The company reiterated that it has a zero-tolerance policy towards child sexual exploitation and abuse material.
According to Meta, it deploys advanced detection technology to identify violating advertisements and remove them quickly. It also said criminals constantly evolve their tactics to bypass automated detection. Following the publication of the BBC investigation, Meta stated it removed the advertisements identified by reporters and denied deliberately targeting users with exploitative content. The company also acknowledged that no automated review system catches every policy violation.
That admission is significant.
It reflects a broader industry challenge. Artificial intelligence can review enormous volumes of content. However, sophisticated bad actors increasingly learn how to manipulate language, imagery and symbols to avoid detection. The result becomes an ongoing technological arms race between platforms and criminals.
A Former Insider Raises Different Questions
The BBC investigation also revisited criticism from a former Meta employee who previously worked on child safety issues. He argued that the company had gradually accepted trade-offs between business priorities and user safety. According to him, discussions increasingly involved balancing revenue with moderation decisions.
He ultimately left the company because he believed user protection was not receiving sufficient priority. Meta has consistently rejected these claims.
The company maintains that protecting children remains one of its highest priorities and that it continuously invests in detection technology, specialised review teams and law enforcement partnerships. Still, the testimony adds another dimension to the debate. When moderation failures occur, critics inevitably question whether platforms invest enough resources into preventing them before they reach users.
India’s Notice Goes Beyond One Advertisement
The government’s response suggests regulators are treating this as more than an isolated moderation failure. According to government sources, MeitY has directed Meta to disable all advertisements and content facilitating access to child sexual exploitative and abusive material.
Officials have also demanded a detailed explanation within seven days.
The notice reportedly follows direct instructions from IT Minister Ashwini Vaishnaw.
The ministry is expected to review Meta’s response before deciding whether additional action is required. The notice also reinforces India’s existing legal framework.
Under the Information Technology Act, publishing or transmitting child sexual abuse material carries stringent criminal penalties. The Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, further require intermediaries to remove such content within twenty-four hours after receiving valid notice. Failure to exercise due diligence may result in intermediaries losing statutory protections available under Indian law.
This Isn’t Just About Meta
The implications extend across the entire creator economy. Instagram today is not simply a social network. It is one of India’s largest advertising platforms. Millions of creators, brands and small businesses rely on Meta’s advertising infrastructure to reach audiences every day.
That relationship depends on trust. Advertisers expect their campaigns to appear in safe environments. Creators expect platforms to protect communities. Users expect illegal content never to receive paid distribution. When that trust weakens, the consequences extend beyond one investigation. Brands become more cautious. Regulators increase oversight. Users begin questioning recommendation systems they cannot see or control.
The Algorithm Is Now Under Scrutiny
The BBC investigation also revives another ongoing debate surrounding recommendation algorithms. The journalists claimed they did not actively search for exploitative content. Instead, the recommendation system allegedly guided them toward increasingly explicit material over time. Meta denies deliberately promoting such content. However, recommendation systems optimise for engagement, not intent.
That distinction matters. If harmful actors learn how recommendation systems behave, even temporary moderation failures can dramatically increase visibility before detection occurs. The issue therefore becomes larger than content moderation alone. It becomes a question of algorithmic responsibility.
What Happens Next?
Meta now faces two simultaneous challenges. First, it must satisfy Indian regulators by explaining how these advertisements allegedly bypassed multiple moderation layers. Second, it must reassure users, advertisers and creators that its advertising systems remain trustworthy. The company’s response will likely shape not only regulatory discussions in India but also wider global conversations around platform accountability.
As governments increasingly demand transparency from technology companies, removing harmful content may no longer be enough. Platforms may also be expected to explain why their systems allowed that content to appear in the first place. That may ultimately become the biggest question emerging from this investigation, not whether illegal content exists online, but whether the systems designed to prevent it are keeping pace with those determined to exploit them.