The Ministry of Electronics and Information Technology (MeitY) has reportedly told Meta to revamp its algorithms and provide a compliance roadmap to fight deepfakes, propaganda and other manipulated or illegal content on its platforms.
According to ANI sources, officials also asked Meta to tighten content moderation, act faster on takedown requests from authorised agencies, follow data-localisation requirements and work more closely with the government to speed up detection and removal of deepfakes, morphed images and misleading posts.
Also Read: Govt-Meta meeting enters technical round on day 3; company lists ways to curb deepfakes, CSAM
The directive came after recent talks in which Meta admitted gaps in its current systems and laid out technical steps to tackle deepfakes and synthetic content.
In a Friday meeting with MeitY officials, Meta's technical team walked through how it currently handles harmful content online. Officials pushed for specific algorithmic changes to limit the spread of propaganda, deepfakes and manipulated media.
Both sides agreed to continue discussions, with another meeting expected next week.
The government is also examining whether Meta complies with intermediary provisions under Indian law or violates them, sources told PTI.
The government has asked Meta to act on deepfakes, child sexual abuse material (CSAM) and unlabelled synthetically generated content. A legal opinion on further action will be taken after reviewing the outcome of the meetings, the sources said.
The government is also likely to call other platforms as it examines whether they meet the definition of intermediaries under Indian law, sources said.
If Meta and other platforms decide what content is shown, it is akin to "publishing" and they must take responsibility for their actions, government sources said.
The latest discussions follow three days of meetings between government officials and Meta's global team, led by the company's global affairs head Joel Kaplan.
After two days of intensive questioning, Friday's discussions focused on technical aspects, during which Meta explained how its systems work and how it plans to address the government's concerns over deepfakes, CSAM and bot accounts.
A key issue discussed was the reappearance, continued circulation and virality of AI-generated harmful content even after it has been flagged. The government also questioned why AI-generated videos without labels continue to remain visible on Meta's platforms despite IT rules requiring synthetic content to be identified and labelled.
Also Read: Meta apology victory for India's laws, Constitution and Parliament: Dubey
The government asked Meta to have greater human oversight in content moderation, with enhanced awareness and understanding of Indian languages as well as local nuances.
The government has asked Meta to take specific measures and report back on the outcomes. Officials will regularly follow up with the company on its progress, sources said.
The government has maintained that the objective is not censorship but to ensure that the company complies with Indian laws.
The meetings followed the temporary restriction of a Facebook post by Prime Minister Narendra Modi. Meta later restored the post, attributing its removal to an error by AI-powered automated content filters, and apologised.
The government had also questioned Meta on its content recommendation algorithms and guardrails to ensure constitutional and legal requirements are being met.
IT Minister Ashwini Vaishnaw had asked Meta whether the company accepted payments to promote content that could trigger public disorder, according to government sources. The government also read out provisions of Section 79 of the Information Technology Act to Meta officials, making it clear that non-compliance with intermediary obligations jeopardises the company's safe harbour protection.
The government had last month served a notice to Meta over alleged child sexual abuse material in paid advertisements on Instagram.
A Parliamentary Standing Committee had also sought an apology from Meta CEO Mark Zuckerberg over the removal of Modi's Facebook post, warning that the protection and immunity available could be withdrawn if no apology was forthcoming within three days.
(With inputs from agencies)
