India's IT Ministry Holds Multi-Day Talks With Meta on Algorithms, Content Moderation and Compliance
The Facts
- MeitY held a third round of discussions with Meta on Friday, shifting from two days of questioning by IT Minister Ashwini Vaishnaw to technical-level talks on the company's platform algorithms and systems.
- The engagement followed the temporary restriction of Prime Minister Narendra Modi's Facebook Reel on July 28, which the Hindustan Times reported left the post inaccessible for around five hours.
- Meta's chief global affairs officer Joel Kaplan met Vaishnaw and apologised on behalf of the company for the restriction on Modi's post.
- The government has asked Meta to take specific measures on deepfakes, child sexual abuse material and unlabelled synthetically generated content, and to report back on the outcomes, with officials saying they will follow up on progress.
- Officials questioned why AI-generated videos without labels remain visible on Meta's platforms despite IT rules requiring synthetic content to be identified and labelled, and raised the continued circulation of harmful AI-generated content after it has been flagged.
- Government sources say the central question under examination is whether Meta's recommendation systems and paid content promotion are compatible with its status as an intermediary under the IT Act, or amount to publishing by deciding "what content is shown to whom."
- Section 79 of the IT Act grants intermediaries "safe harbour" from liability for third-party content, conditional on compliance with due diligence requirements; losing it would make Meta liable for user-generated content in India.
- A parliamentary panel sought an apology from Mark Zuckerberg over the restriction of Modi's post, warning that it would recommend withdrawing Meta's safe harbour immunity if he did not comply.
- The government is awaiting a legal opinion on further action after reviewing the outcome of its meetings, and is likely to seek responses from other social media platforms on whether they meet the definition of intermediaries.
- Meta acknowledged shortcomings during the meetings, with sources saying the company admitted gaps in its algorithmic mechanisms and expressed regret over lapses raised by officials.
Context
What is "safe harbour" and why is it central to this story?
Section 79 of India's IT Act gives online intermediaries immunity from liability for content posted by their users, but that protection is conditional on the platform meeting due diligence obligations Rediff.com India Lt…,mint. Government sources say they are examining whether Meta's recommendation algorithms and paid promotion of content mean it is effectively acting as a publisher rather than a neutral intermediary India Today,mint,Economic Times. If the protection were withdrawn, Meta could be held legally responsible for user-generated content on Facebook, Instagram and WhatsApp in India CNBC,mint.
What specific changes has the government asked Meta to make?
MeitY has asked Meta to submit a compliance roadmap covering steps to strengthen content moderation and curb the spread of deepfakes, propaganda and other unlawful material, including algorithmic changes India Today. Officials also pressed for greater transparency into how Instagram and Facebook algorithms determine reach and visibility, and for more cooperation with Indian law enforcement on CSAM and online fraud Indian Express. Separately, sources told NDTV the government asked Meta to increase human oversight in content removal and to add reviewers fluent in Indian languages and familiar with local context NDTV.
What else is Meta facing at the same time?
In the United States, a New Mexico court ordered Meta on Thursday to pay $567 million over harms to young people linked to its platforms, with $420 million allocated to treatment services and the remainder to prevention, awareness and screening programmes over five years India TV News,CNBC. The ruling followed an earlier trial phase in which a jury ordered a $375 million payment India TV News.
Where Left and Right agree, and where they split
- Where Left and Right agree
- Unlabelled AI video persists despite rules requiring it be identified, and safe harbour is a conditional bargain — so pressing whether algorithmic curation is publishing is legitimate.
- Where Left and Right split
- Whether the story is about children and deepfake targets bearing the cost of unenforced labelling rules, or about a platform that curates forfeiting its claim to mere intermediation.
How left and right read it
Officials say unlabelled AI-generated video stays visible despite rules requiring synthetic content to be labelled, and that harmful AI content keeps circulating even after it is flagged — failures borne by children and by people targeted through deepfakes. That deserves the insistence evident in the apology offered for the restriction of the Prime Minister's post and the panel's warning about safe harbour. Whether algorithms amount to publishing should be pressed on behalf of those with no such leverage.
MeitY moved to technical-level talks on Meta's algorithms Friday, after officials asked why unlabelled AI-generated videos remain visible despite IT rules requiring synthetic content to be identified. Safe harbour under Section 79 was never a grant; it is conditional on due diligence, and the bargain runs both ways. Now the harder question is on the table. If recommendation systems and paid promotion decide what content is shown to whom, on what principle does that remain mere intermediation?
“According to the sources, if a platform determines what content gets shown to users, it amounts to publishing, and platforms would then have to take responsibility for their actions” — The Telegraph
The receipts — all 68 sources
Independent coverage (50)
Facts first. Then every angle.
The day’s biggest stories in one short brief — the facts everyone agrees on, then the competing values behind the headlines. Free in your inbox.