From Passive Channel to Active Watchdog: Revisiting Safe Harbour under India’s 2026 IT Rules

India’s 2026 IT Rules move intermediaries towards more active duties. This article examines synthetic-content labelling, compressed takedown periods and the danger of over-removal.

I. Introduction

Section 79 of the Information Technology Act gives intermediaries conditional protection for third-party information when statutory requirements are met. In Shreya Singhal v. Union of India, the Supreme Court confined the actual-knowledge route for removal to a court order or appropriate government notification. The amendments notified on 10 February 2026 add rules for synthetically generated information and compress certain removal periods. They raise a practical question: can an intermediary comply quickly while still respecting lawful speech?

The amendments move beyond a purely reactive model by adding technical duties for services that enable synthetic-content creation. This article asks how those duties and shorter removal periods can be applied consistently with Section 79 and the protection of lawful expression.

II. The Regulatory Framework: A Critical Analysis of What's Different

Rule 2(1)(wa) defines synthetically generated information by reference to realistic computer-created or altered audio, visual or audio-visual material. It excludes specified routine edits and non-deceptive processing. Rule 3(3) applies additional technical and labelling duties to intermediaries that enable SGI creation or alteration, with provenance information required to the extent technically feasible. Rule 4(1A) adds user declaration and verification duties for significant social-media intermediaries. These duties do not convert every intermediary into a general monitor of all speech.

Rule 3(1)(d) requires removal or disabling access to unlawful information within three hours after a court order or a reasoned written intimation from an authorised government officer. The obligation concerns unlawful information generally, rather than SGI alone. Rule 3(2)(b) provides a separate two-hour route for complaints about specified private, nude, sexual or impersonating material, including certain morphed images. Other grievance time limits also changed. Under Rule 7, a failure to observe due diligence may affect the intermediary’s Section 79 protection; it does not, by itself, create a seven-year offence under Section 69A.

III. Doctrine of Dissonance: Due Diligence vs De facto Pre-Censorship?

The shorter order-compliance period needs to be read against Shreya Singhal. The Court’s actual-knowledge threshold guards against removal based only on an untested private allegation. A valid court order or authorised government intimation still provides that legal trigger under Rule 3(1)(d). Yet a three-hour period may leave little room to check the scope of an unclear notice, identify the specific material and limit removal to what the law requires. The concern is a practical incentive to remove more speech than the order calls for.

The SGI duties also raise a separate question about proactive technical measures. Platforms that enable generation must take reasonable steps against specified unlawful material and label lawful SGI; their obligations should be applied with care to avoid treating harmless edits, satire or political expression as illegal. The legal distinction between technical prevention duties and removal after an authorised order is crucial to the balance drawn in Shreya Singhal.

Supporters of short deadlines point to the speed at which non-consensual intimate images and impersonating content can spread. Prompt action can reduce harm. But quick removal and accurate decisions work best when notices identify the exact content and legal ground, and when a fast review process can correct mistakes. Comparisons with other jurisdictions must account for differences in the content covered and the procedural safeguards.

IV. Procedural Safeguards: Sufficient or Ornamental?

The grievance and appeal mechanisms permit a user to challenge an intermediary decision, but their time frames are much longer than the two- or three-hour periods for initial action. For a news report or time-sensitive political comment, later restoration may not fully cure an erroneous removal. A prompt review route and clear reasons for action would make the initial speed less likely to suppress lawful expression.

This concern extends beyond synthetic media because Rule 3(1)(d) addresses unlawful information generally. The two-hour complaint route is narrower, while SGI labelling and technical duties apply under different conditions. A workable approach should distinguish clearly identifiable intimate or impersonating harm from contested expression that requires more context before a decision is made.

V.Conclusion

The amendments increase intermediaries’ responsibilities through shorter response periods and additional SGI duties. Their legal effects depend on the particular trigger: a court order or authorised written intimation under Rule 3(1)(d), a specified complaint under Rule 3(2)(b), or a technical duty under Rule 3(3). Non-compliance may affect Section 79 protection, but it does not automatically impose the criminal penalty in Section 69A. Clear notices, proportionate technical measures and rapid correction of mistaken removals would help reconcile the rules with free expression.

Primary sources

MeitY: consolidated intermediary rules as amended in February 2026

Information Technology Act, 2000

Technology LawDigital RightsIntermediary Liability