Q-Cyber Magazine

AI · 8 April 2026 · 4 min read

Deepfakes and India's 2026 rules on synthetically generated information

Amendments to the IT Rules, notified on 10 February 2026 and in force from 20 February 2026, require intermediaries to prevent unlawful synthetic media and to label permissible synthetically generated information.

Abstract graphic of a network of connected nodes in Knowlcom blue

Synthetic media is no longer only a research problem. In February 2026, India amended the Information Technology Rules, 2021, to set due-diligence duties for synthetically generated information, including deepfakes.

What the amendment requires

The Ministry of Electronics and Information Technology has been explicit that synthetically generated information can be useful — for language, education, accessibility and public services — and that the new duties are aimed at misuse, including deepfakes, misinformation and other unlawful content. A new Rule 3(3) sets a dedicated due-diligence framework: measures to prevent prohibited synthetic information, mandatory labelling and provenance for what is permitted, and safeguards against the removal or tampering of those labels and identifiers.

The labelling rule is practical. Visual or audiovisual content must carry a prominent on-screen label or notice identifying it as synthetically generated. Audio must carry a prominent prefixed audio disclosure. Where technically feasible, the content should also carry permanent metadata or a technical provenance mechanism, including a unique identifier, so that the information can be recognised as synthetic and linked to the computer resource that created or altered it.

The Press Information Bureau's 25 March 2026 account, based on a Lok Sabha reply, adds that platforms must also inform users of the legal consequences of unlawful synthetic content, and that significant social media intermediaries carry additional duties. Unlawful categories expressly include content that is obscene, misleading, impersonating, or harmful to children, as well as non-consensual intimate imagery.

Exhibit 1

Public steps toward India's synthetic-media due diligence

  1. 15 March 2024

    MeitY advisory to intermediaries on due diligence, including risks from tools that create or modify text, audio and audiovisual content

  2. 11 November 2025

    Standard operating procedure released to curtail dissemination of non-consensual intimate imagery

  3. 10 February 2026

    IT Rules Amendment on synthetically generated information notified (G.S.R. 120(E))

  4. 20 February 2026

    Amendment Rules come into force

  5. Within 3 hours

    Removal or disablement after a court order or a reasoned intimation from the authorised government officer

Source: Press Information Bureau, 25 March 2026 (PRID 2245053); MeitY, FAQs on the IT Rules Amendment, 10 February 2026, including the summary table of key timelines.

Implications beyond social platforms

Any organisation that generates or hosts realistic audio or video — in advertising, customer service, training, or political communication — should assume that labelling and provenance will become a normal control, not an exception for large platforms. Enterprises that use generative models to produce public-facing media should decide, in advance, how a "synthetically generated" notice will appear, how metadata will be preserved, and how a takedown request will be met inside three hours when the law requires it.

The amendment does not, by itself, solve authentication of real footage. It creates a duty to mark what is synthetic and to prevent what is unlawful. Detection, provenance from the camera onward, and user awareness remain complementary work.

Implications

  1. Inventory public-facing synthetic media and apply prominent labels plus, where feasible, durable provenance.
  2. Protect labels from being stripped, in line with the Rule 3(3) anti-tampering duty.
  3. Rehearse the three-hour removal path for court or authorised-government intimations.
  4. Treat impersonation and non-consensual imagery as prohibited classes, not as edge cases.

What this means for AI

The legal baseline in India is now clear enough to design against: prevent unlawful synthetic media, label what is permitted, and be able to take it down quickly when required. Organisations that build those controls into generation and hosting pipelines will be faster, and safer, than those that treat labelling as a later overlay.

Sources

  1. PIB: Government strengthened IT Rules to address synthetically generated information
  2. MeitY FAQs on the IT Rules Amendment, 2026
  3. MeitY document list for the IT Rules, 2021, including the 10 February 2026 amendment