Newsletter
Meta is facing fresh scrutiny in India after the government ordered the company to immediately remove Instagram advertisements and content that promote or facilitate child sexual abuse material (CSAM).
It has also given the company seven days to explain how such advertisements were approved in the first place. The action follows a BBC Eye investigation, which found that a newly created Instagram account in India was shown paid advertisements that allegedly directed users to Telegram channels where child sexual abuse material was being sold.
This is not about a random post uploaded by a user. These were paid advertisements. Someone reportedly bought advertising space on Instagram, Meta's systems approved those ads, and they were shown to users. That distinction changes the entire conversation because it raises questions not just about content moderation, but also about how one of the world's largest advertising businesses reviews and approves millions of advertisements every day.
Let’s first understand how Meta's advertising system works.
Every day, millions of businesses submit advertisements across Facebook and Instagram. Before an advertisement goes live, it passes through an automated review process that uses artificial intelligence to check whether it violates Meta's advertising policies. Depending on the content, some advertisements may also undergo human review. The company has repeatedly said that it has a zero tolerance policy for child sexual exploitation and uses advanced AI systems to proactively detect such material.
Yet, according to the BBC investigation, advertisements promoting child sexual abuse material still managed to pass through these filters. If that finding is accurate, it exposes a worrying gap in a system that processes an enormous volume of advertisements every single day.
This is also not happening in a legal vacuum. India already has some of the world's strictest laws against online child sexual abuse material. Section 67B of the Information Technology Act makes it illegal to publish, transmit, browse, advertise, promote or distribute sexually explicit content involving children. A first conviction can attract up to five years in prison along with a fine of up to ₹10 lakh, while repeat offences can result in up to seven years in prison.
Beyond criminal law, India's Information Technology Rules, 2021 place additional responsibilities on large social media platforms. Companies designated as Significant Social Media Intermediaries are expected to use automated tools to proactively identify content related to rape, child sexual abuse and other illegal material. They are also required to establish grievance mechanisms, remove unlawful content when notified and demonstrate due diligence in preventing such content from spreading.
The challenge, however, is that moderating content at internet scale is far more difficult than it sounds. Meta says its family of apps serves around 3.5 billion people every month. Every minute, users upload photos, videos, messages and advertisements in hundreds of languages. Artificial intelligence has become the company's first line of defence because manually reviewing everything would be practically impossible.
But AI is not perfect. Criminals constantly change keywords, images, spellings and symbols to avoid automated detection. Sometimes they hide illegal activity behind seemingly harmless advertisements or redirect users to encrypted messaging platforms where transactions happen away from public view. That cat-and-mouse game has become one of the biggest challenges for social media companies worldwide.
The numbers show just how large the problem has become. The US-based National Center for Missing and Exploited Children, better known as NCMEC, received 20.5 million CyberTipline reports related to suspected online child sexual exploitation in 2024. While that was lower than the previous year's figure because of changes in reporting methodology by some platforms, it still represents tens of millions of suspected incidents being flagged in a single year.
The organisation has also warned that generative AI is rapidly making the situation worse. Reports involving AI-generated child sexual abuse material have increased sharply as image generation tools become more sophisticated and widely available.
Meta itself has acknowledged the scale of the problem. According to its transparency reports, Instagram removed millions of pieces of child nudity and child exploitation content during 2025. The company also disabled more than 635,000 Facebook and Instagram accounts linked to child safety violations, including over 135,000 Instagram accounts that allegedly left sexualised comments or solicited inappropriate images involving children.
India is also not alone in tightening oversight of social media companies. Governments across the world have increasingly shifted from asking platforms to moderate harmful content voluntarily to making them legally responsible for preventing it. The European Union's Digital Services Act requires large online platforms to assess and reduce systemic risks, especially those affecting children. The law also pushes companies to strengthen age verification, improve transparency and explain how their recommendation systems work.
The United Kingdom has taken a similar route through the Online Safety Act, which places legal duties on technology companies to tackle illegal content, including child sexual exploitation. Companies that fail to meet these obligations can face significant penalties. Australia has gone even further by giving its eSafety Commissioner the authority to demand detailed information from platforms about how they detect and remove child sexual abuse material. Regulators there have already used those powers against major technology companies when they believed compliance was inadequate.
These developments point to a much broader shift in internet regulation. Governments are becoming less willing to accept that argument. The focus is now moving towards platform accountability, especially when recommendation algorithms, advertising systems or automated tools actively amplify or distribute harmful content.
The Indian government's notice, therefore, is about much more than a handful of advertisements. It reflects a growing belief among regulators that platforms should not only remove illegal content after it is discovered but should also be able to explain why it was allowed to appear in the first place. As artificial intelligence makes both content creation and content moderation more complex, that question is likely to become one of the defining challenges for every major social media platform, not just Meta.
