New Delhi: Meta has told the Home Ministry’s cybercrime wing that it acted on more than 61,000 “violent and inciteful” Facebook and Instagram posts in the two weeks between 25 July and 7 August, that objectionable AI content “is presently not being recommended on the platform”, and that it had “significantly strengthened” its safeguards against adult nudity and sexual content.
The fortnight Meta has cited came right after the most volatile stretch of the Cockroach Janta Party (CJP) protests over the NEET paper leak issue. The youth-led movement’s ‘Chalo Sansad’ march on 20 July was met with tear gas and lathi charge, and the CJP had called nationwide protests on 24 July against “police brutality”.
The figures were shared at a 21 August meeting with the Indian Cyber Crime Coordination Centre (I4C), chaired by its CEO Rajesh Kumar. The minutes of this meeting were filed in the Supreme Court on 7 October with the Centre’s status report in the suo motu case on digital arrest scams, which ThePrint has seen.
The minutes do not say what the 61,000 “violent and inciteful” Facebook and Instagram posts were related to.
Meta also said it had acted on more than 7.5 million pieces of AI-generated content between March and 18 August, including over 3 million in that same fortnight of 25 July and 7 August. Such content on its platforms, it said, had declined by about 95 percent.
The minutes describe the 21 August meeting as a follow-up to reviews by the Union Home Secretary, by the Parliamentary Standing Committee on Communications and Information Technology, and by the court-monitored committee on ‘digital arrests’. It came 18 days after the parliamentary panel’s stormy session with Meta, where members had told Meta that an apology would not suffice when it came to the brief removal of Prime Minister Narendra Modi’s 23 July video message to the youth from his Facebook page, and had raised questions about algorithmic amplification.
I4C was unimpressed by the totals. It asked Meta for a note on the “70 lakh” pieces removed, with specific examples and direct links, “rather than just aggregate numbers”.
Meta was also told to brief I4C in person on its “core algorithmic principles”, addressing concerns over “how the algorithm amplifies or suppresses specific content and viral accounts”. It was asked for “a detailed analysis and findings regarding coordinated bot activity and shared bot networks across prominent handles”.
Meta was also told to send its draft community guidelines “for review”, making sure they cover deepfakes, “pornographic material involving prominent personalities” and content that is taken down and re-uploaded.
I4C also asked Meta to explain why it reports child sexual abuse material to “third-party international agencies”—the US-based National Center for Missing and Exploited Children—rather than directly to Indian authorities.
The August meeting capped months of tightening oversight. In May, Meta agreed to name dedicated points of contact for I4C and to meet it “initially on a weekly basis”. It also agreed to compile a cumulative report on the scam networks it had disrupted, to be reviewed after six months.
On scams, Meta cited action against more than 9.89 lakh ads and ad accounts targeting Indian users, and said 9,400 WhatsApp accounts were banned over digital arrest scams. However, I4C said that “the extent of enforcement presently undertaken remains insufficient”. Such scams, it added, strip “innocent and vulnerable sections of the population of their lifetime savings and investments”.
Earlier minutes show the friction building. In April, I4C told Meta that investment scam ads, “chiefly” on Facebook, were causing losses of about $2 million a day in certain categories. The ads targeted people aged 55 to 65 because of the “higher probability of defrauding accumulated lifetime corpus”. Many impersonated the Prime Minister and the Finance Minister, and were paid for by foreign entities, including from Hong Kong. Threat reports sent to Meta, I4C said, had gone unanswered for “multiple weeks/months”.
In May, I4C said eight China-based advertisers it had repeatedly flagged had “continued to receive advertisement credits”, and that scam ads were running through verified, blue-tick accounts. It asked Meta to share intelligence on scam compounds with Indian agencies “in line with what Meta has been doing with the FBI”. Meta said the proposal “could be examined further”.
Meta also pushed back on a “kill switch” to automatically cut off ‘digital arrest’ scam calls. Instead, it offered its own fixes for WhatsApp—showing a caller’s location and account age, hiding profile photos from unknown callers, and adding warnings during long video calls. I4C asked Meta for “detailed reasoning” behind its objections.
On 29 October, the Supreme Court had directed WhatsApp and other platforms to submit their own plans to stop digital arrest scams. The case will be heard next on 23 November.
(Edited by Gitanjali Das)