Meta must follow Indian law, not just global rules, say government sources after MeitY talks
New Delhi, 9 August
Meta's platforms must be governed by Indian law, not merely by the company's global policies, government sources clarified on Sunday and noted that it would be factually inaccurate to say that the government had "directed Meta to completely overhaul its algorithm."
While certain concerns remain, officials said the core intent of recent engagements with Meta is to ensure the platform's guidelines and content policies are fully aligned with Indian law.
Sources further pointed out that India's cultural context is often not fully understood by teams based outside the country, and that Meta needs greater local input including a stronger grasp of Indian languages and cultural nuance to moderate content effectively.
On the issue of Child Sexual Abuse Material (CSAM), government sources were unequivocal, stating there is zero tolerance and zero compromise on such content. Allowing such material to remain on the platform amounts to a serious violation of rules, they said, adding that Meta must adopt stronger, more proactive measures, with concrete action expected rather than assurances alone.
Referring to deepfakes involving prominent personalities, sources said content posted from authorised or verified handles should not be misidentified as deepfakes. They noted that such sensitive cases cannot be left entirely to automated detection systems, and called for a "human-in-the-loop" mechanism to ensure proper review before any takedown action.
On synthetically generated content, officials raised concerns over repeated reappearance of material already identified and removed as deepfakes. Meta has been asked to specify what concrete steps it is taking to prevent such content from resurfacing after removal.
The government also sought clarity on Meta's recommendation and virality algorithms, questioning why users are frequently shown content unrelated to their stated interests. Sources said this points to underlying flaws in the recommendation system that need to be understood and addressed. Meta was specifically asked why synthetically generated content continues to be amplified by its algorithm, with officials insisting that once such content is flagged, it should neither be recommended further nor allowed to resurface repeatedly.
The clarifications came after a meeting on Friday between Meta's technical team and officials from the Ministry of Electronics and Information Technology (MeitY), during which the company's representatives walked through existing systems for handling harmful content online. Technical discussions between the two sides are ongoing, with both parties agreeing to continue talks. Another round of meetings is expected in the coming week.
— ANI
Reader Comments
The "human-in-the-loop" mechanism is crucial! Automated systems alone cannot understand context, especially for something as serious as deepfakes of public figures. A single wrong takedown can ruin someone's reputation. Glad MeitY is having these technical discussions. Better late than never!
Honestly, I'm impressed with the specific issues raised here - from language nuance to algorithm transparency. It's not just about censorship; it's about accountability. The point about verified handles being misidentified as deepfakes is particularly important. Need more of this technical rigor in policy-making. 👍
As someone who follows tech policy globally, this is a significant shift. India isn't just asking Meta to follow rules - they're asking for algorithmic transparency, which even Western regulators have struggled with. The question about content resurfacing after removal is spot on. Hopefully this leads to stronger platform accountability, not just in India but worldwide.
The cultural context point is so true! I've seen many posts in Tamil and Hindi that get wrongly flagged because the moderation team in California doesn't understand local slang or cultural references. If Meta wants to operate in India, they must hire local experts. It's about respect for our languages too.
A balanced approach from Indian authorities here. They're not insisting on blanket censorship but demanding proper scrutiny where it matters - especially CSAM and deepfakes. The mention of "concrete action rather than assurances" should give Meta pause. Actions speak louder than promises. 📢
We welcome thoughtful discussions from our readers. Please keep comments respectful and on-topic.