The government has summoned Meta’s global team for two days of talks beginning on August 5, expanding its scrutiny of the technology company to child safety, artificial intelligence, account restrictions and content moderation.
Officials are expected to question the Facebook and Instagram parent over the handling of child sexual exploitative and abuse material, including allegations that paid advertisements on Instagram promoted access to illegal content. The discussions on August 5 and 6 will also examine the spread of AI-generated material and safeguards governing prominent verified accounts.
Electronics and Information Technology Secretary S Krishnan confirmed that the issues would be raised during the global team’s visit. The meeting marks a widening of the government’s engagement with Meta beyond an earlier dispute involving the temporary restriction of a Facebook post by Prime Minister Narendra Modi.
The post was briefly limited by Facebook’s automated systems before being restored. Meta apologised for the action, but officials have sought a fuller explanation of how the decision was made and whether algorithmic processes can incorrectly restrict authentic material posted by verified public figures.
The government is expected to ask for technical details about Meta’s moderation architecture, escalation procedures and human-review mechanisms. A technical representative from Instagram has also been sought, reflecting demands for clarity on how advertisements, recommendations and account-level enforcement decisions are assessed.
Concerns over child abuse material intensified after allegations that advertisements on Instagram directed users towards outside channels offering exploitative content. The government subsequently issued a notice seeking information on Meta’s preventive systems, the removal of offending advertisements and action taken against the accounts behind them.
Meta has said it maintains zero tolerance for child exploitation and uses artificial intelligence, behavioural signals and automated detection systems to identify suspicious activity. The company said it removed more than four million suspicious accounts worldwide last year and 160,000 accounts in the country during a six-month period.
It also reported removing 36 million pieces of content linked to child exploitation through automated systems. Meta has argued that several advertisements and accounts connected to the allegations had already been disabled before the matter was brought to its attention, with further action taken after an internal investigation.
Officials are likely to examine whether those measures were fast and comprehensive enough, particularly when paid advertising systems are involved. The appearance of prohibited promotions raises questions about how advertisers pass automated checks, whether commercial incentives weaken enforcement and how recommendation systems may amplify harmful networks.
The talks are also expected to cover generative AI and synthetic media. Revised intermediary rules that took effect in February strengthened duties concerning AI-generated audio, video and images. The framework requires platforms to take reasonable technical measures against unlawful synthetic material and provide prominent labelling or identifiers for permissible AI-generated content.
Those requirements have gained importance as generative tools make it easier to produce convincing impersonations, deceptive political material and synthetic sexual imagery. Child-protection specialists have warned that AI-generated abuse material can revictimise survivors, imitate identifiable children and overwhelm investigators attempting to distinguish synthetic files from evidence of physical offences.
Government representatives are expected to seek assurances that Meta’s AI systems can recognise prohibited synthetic material without indiscriminately restricting lawful content. The challenge is particularly significant for platforms operating at national scale, where automated moderation must interpret multiple languages, dialects, cultural contexts and political references.
The temporary restriction of Modi’s post has placed algorithmic accountability at the centre of the dispute. Officials want to establish whether Meta’s systems treated authentic political communication as manipulated material and what safeguards exist to prevent similar errors involving constitutional authorities, verified journalists or public institutions.
Meta faces a parallel challenge in explaining how it balances automated enforcement with human oversight. Automated tools can act rapidly across billions of posts, but errors can become politically sensitive when they affect influential accounts. Excessive reliance on human review, however, can slow action against abusive content that spreads quickly.
The company’s representatives may also be questioned about the suspension or restriction of other prominent verified accounts and the availability of effective appeal channels. Authorities are seeking more predictable procedures, particularly where platform decisions may affect public communication, access to information or public order.
