For a company that reaches nearly a billion Indians, Meta has spent an oddly large amount of time this year saying sorry. The latest apology, according to some reports, sent by Mark Zuckerberg himself to India’s parliamentary panel and the government, is the most striking yet. It followed a three-day ultimatum, a summons of Meta’s most senior global officials, and a warning that the company could lose the one legal shield that makes its business in India possible. When a firm of Meta’s size is made to apologise under the threat of losing its “safe harbour”, it is worth pausing to ask what exactly went wrong, and whether “we’re sorry” is going to be enough.
The trigger was small in appearance but large in symbolism. On July 23, Prime Minister Narendra Modi posted a video addressed to students and young voters, promising strict action against paper leaks and speaking about exam reforms. For roughly five to six hours, Facebook took that video down. Meta later called it an operational error. But in a country still raw from the NEET paper-leak protests, the timing could not have been worse. A Prime Minister speaking directly to anxious students, silenced by a foreign platform’s automated system, even briefly, was never going to be brushed aside as a technical hiccup.
The video, though, was only the spark. Around it sat a pile of dry tinder. During the protests at Jantar Mantar, a wave of AI-morphed and deepfake videos began circulating on Facebook and Instagram. Some falsely showed Union Minister Piyush Goyal making violent remarks; others put words in the mouths of BJP leaders. Fact-checkers and the government’s own Press Information Bureau flagged several of these as fabricated. Police in Hyderabad and Maharashtra have since registered FIRs, and, remarkably, named Meta’s India head in some of them, arguing the platform failed to detect or curb the morphed content fast enough. Add to this the government’s charge that Meta’s ad-review system let child sexual abuse material slip into paid Instagram distribution, and you begin to see why the parliamentary committee, chaired by BJP MP Nishikant Dubey, was in no mood to accept vague assurances.
When Meta’s Chief Global Affairs Officer and its global policy head were summoned, the questioning went beyond the Modi video. Lawmakers pressed them on algorithmic bias, on why the same automated systems both remove content and decide the appeals against those removals, and on the platform’s role in public order during a tense political moment. The committee’s message was blunt: Comply, apologise unconditionally, or risk being treated not as a neutral intermediary but as a publisher, which would strip away the immunity granted under Section 79 of the IT Act. That is the real story here. It is not one deleted video. It is a government arguing that a platform which decides who sees what has stopped being a passive pipe and has become a publisher responsible for what flows through it.
This brings me to my first analytical point: Companies like Meta can no longer treat regulation and the law as an afterthought to be managed by a local policy team and a well-worded tweet. For years, Big Tech operated on a simple assumption: That it was too large, too essential, and too fast-moving for any single government to seriously discipline. India is dismantling that assumption. The threat to withdraw safe harbour is not symbolic. Lose intermediary status, and every piece of user content becomes a potential legal liability, in a market of hundreds of millions. That is an existential risk, not a public-relations one.
Meta’s recent pattern, from the false claim about India’s 2024 election result to the CSAM lapses to the Modi video, shows a company repeatedly reacting after the damage is done, then apologising. But an apology is a bandage, not a cure. What regulators want, and what the law increasingly demands, is provable compliance built into the system beforehand: Faster grievance redressal, transparent moderation, clear labelling of AI content, and a human being who can be held accountable when the machine gets it wrong. Firms that keep outsourcing judgment to algorithms while laying off the very staff who once handled appeals are, frankly, asking for exactly this kind of confrontation. Pulling up their socks means investing in legal and regulatory readiness as a core cost of doing business, not as damage control after the summons arrives.
My second point is about the harder, slower duty: Safety and societal responsibility. A platform is not a neutral noticeboard. Its algorithms decide what spreads, and in a country as diverse and politically charged as India, that power carries real-world consequences. When a fabricated video of a minister calling for violence spreads faster than the truth, the harm is not abstract. It can inflame communities, mislead voters and endanger public order. Meta cannot claim credit for connecting a billion people and then shrug when that same reach is used to circulate deepfakes, scams or abusive material.
Detecting AI-morphed content, protecting children from exploitative ads, and moving quickly during moments of social tension are not optional extras; they are the basic duty of care a platform owes the society it profits from. This is not an argument for heavy-handed censorship, and genuine free-speech concerns must not be waved away, since a system that wrongly deletes a Prime Minister’s video can just as easily silence an ordinary critic. But the answer to bad moderation is better, more transparent moderation, not indifference. Being attentive to societal concerns means designing for the worst-case use of your platform, not just the best-case advertising pitch.
So where does this leave us? Zuckerberg’s apology may well cool tempers for now, as an earlier one did in January. But India has shown its hand, and other governments are watching. The message to Meta and its peers is simple. The era of moving fast, breaking things and saying sorry later is closing. The platforms that survive the next decade will be the ones that treat the law, and the societies they operate in, with the seriousness both deserve, before the algorithm blinks again.
The writer is a defence and tech policy adviser and author of The Digital Decades: On 30 Years of the Internet in India
Share
