What the advisory does
On 8 October 2026, MeitY's Cyber Laws Division sent an advisory to all social media intermediaries, effective immediately. According to MediaNama's report on the advisory, it tells platforms to "identify and appropriately label/contextualise manipulated, synthetically generated or misleading content", particularly where it is presented as depicting a current event. It also asks platforms to give priority to complaints from "authorised Government agencies" about content with potential implications for public order.
The case for acting is real. Old footage recirculated as breaking news, and impersonation of public authorities, can spread faster than any correction during a tense moment. Labelling is also a gentler tool than removal: it leaves the speech up and adds context. If this were a narrow, transparent measure, a pro-speech publication could support it.
But this advisory is neither narrow nor transparent, and it matters because of what it changes in the existing rulebook.
A scope jump from eight months ago
MeitY notified an amendment to the IT Rules, 2021 on 10 February 2026, and it came into force on 20 February. That amendment created the category of "synthetically generated information" (SGI): audio, visual or audio-visual content that appears real. Text-only content falls outside it. The DSCI summary notes the Final Rules followed an October 2025 draft that went through public consultation. MediaNama reports that the notified definition is narrower than that draft, which covered anything that "reasonably appears to be authentic or true". Industry participants had called the draft overbroad.
The notified definition therefore reflected a considered narrowing after consultation. The October advisory's labelling clause carries no such limit, and MediaNama found it does not refer to the February amendment anywhere. In effect, the scope that was cut back through a consultation has returned through a letter. Nothing in the advisory went through public comment.
"Misleading" is not a technical category
SGI labelling is at least tractable. A platform can run detection tools and provenance checks on a video. "Misleading" is different. It covers selective quotation, contested statistics, satire, opinion, and fast-moving claims that are unverified today and confirmed tomorrow. The advisory gives platforms no test for any of this. It does not say how "enhanced diligence" differs from the "due diligence" the Rules already require, and it does not define "authorised Government agencies" or say what "priority" means for their complaints.
The practical result is predictable. A platform facing an immediate-effect directive from the ministry that regulates it, with no defined standard, will over-label and over-remove rather than risk a dispute. Text posts, the main vehicle for political commentary, are exactly what the February Rules deliberately left out. Under this advisory they are back in scope.
The safe-harbour question
The advisory says its aim is to preserve public order "while respecting lawful expression" and that actions must be consistent with the IT Act, 2000 and the IT Rules, 2021. It does not cite a specific provision. That matters because the Supreme Court's reading of intermediary liability in Shreya Singhal v. Union of India (24 March 2015) is the baseline. The Court upheld Section 79(3)(b) only by reading it down: an intermediary has "actual knowledge" when it receives a court order or a notification from the appropriate government or its agency, and that order or notification "must strictly conform to the subject matters laid down in Article 19(2)". The Court reasoned that intermediaries cannot be expected to judge the legitimacy of millions of takedown requests. The judgment text lists "public order" among the Article 19(2) grounds, so the ministry has a constitutional hook for public-order concerns. Whether "misleading" content, as such, falls within Article 19(2) is a separate question the advisory never addresses. Falsity alone is not an enumerated ground.
An advisory is not a notified rule and, on its face, lists no penalty. MediaNama notes that it is unlike MeitY's November 2025 and December 2025 advisories in having no stated consequences for non-compliance. That ambiguity cuts both ways. Platforms may treat it as optional guidance, or they may reasonably fear that ignoring it will be held against their due-diligence record. The ministry's own notification of the February Rules states that safe-harbour consequences attach to due-diligence failures, so a soft instrument can still carry hard leverage.
The timing problem
MediaNama reports the advisory arrived about a week after protests in Delhi over the Chief Election Commissioner and the revision of electoral rolls, though it names no event. Whatever the trigger, a directive that targets "current events" and prioritises government complaints on public order puts the state in the position of shaping what is labelled during political contention. That is the context in which independent fact-checking and platform judgment matter most, and the advisory sidelines both.
A better design
India already has the pieces of a proportionate system. A platform can label content that fails provenance checks, and users can add context through community mechanisms. Courts and notified government channels can order removal within Article 19(2). A proportionate approach to misleading content would:
- limit any labelling mandate to a defined category, such as old media recirculated as current, rather than the open-ended "misleading";
- publish criteria for "authorised Government agencies" and require that their requests be logged and reviewable;
- put the measure through the same consultation that narrowed the SGI definition; and
- give platforms a clear statement that labelling in good faith does not affect safe harbour.
India's February Rules showed that MeitY can listen and narrow a rule. This advisory should be put through the same process.