CLAT-2027 Blog

MeitY Tightens AI Labelling Rules: Deepfakes, Safe Harbour and Article 19 for CLAT 2027

MeitY IT Rules AI labelling deepfakes SGI

CURRENT AFFAIRS | 22 April 2026

CLAT GK + CONSTITUTIONAL LAW (ARTICLE 19) & TECHNOLOGY

The Ministry of Electronics and Information Technology (MeitY) has notified amendments to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 — bringing Synthetically Generated Information (SGI) squarely within the intermediary-liability and due-diligence net. The October draft had prescribed that AI labels occupy at least 10% of a visual display or audio duration; the final notification replaces this with a qualitative “prominent, easily noticeable and adequately perceivable” test.

For CLAT 2027, this is a live controversy that sits at the intersection of Article 19(1)(a) free speech, Article 21 privacy/dignity, and statutory intermediary liability — a triple-barrel exam combination.

Want structured CLAT preparation? Try our free 5-day Bodh Demo Course with live classes and expert guidance. Start Free →

Statutory & Constitutional Framework

  • IT Act, 2000 § 79 — “Safe harbour”: an intermediary is not liable for third-party content if it exercises due diligence and acts on “actual knowledge” of unlawful content.
  • IT Act § 66D — Cheating by personation using a computer resource (deepfakes fraud).
  • IT (Intermediary Guidelines) Rules, 2021 — require grievance officers, takedown within 36 hours, and now SGI labelling.
  • Article 19(1)(a) — Freedom of speech and expression; restrictable only on grounds listed in Article 19(2): sovereignty/integrity, security of State, friendly relations, public order, decency/morality, contempt of court, defamation, incitement to offence.
  • Article 21 — Right to privacy and dignity; umbrella for the right against being deepfaked.
  • DPDP Act, 2023 — Digital Personal Data Protection Act; governs consent-based processing of identifiable personal data.

Why This Matters for CLAT 2027

  • Proportionality is the new Wednesbury: After Puttaswamy (2017) and Anuradha Bhasin (2020), the test for any rights restriction is proportionality — legitimate aim, rational nexus, necessity, balance. Expect a principle-application question.
  • Labelling ≠ prior restraint: A regime that labels rather than prohibits content is prima facie a less-restrictive means — constitutionally safer than an outright ban.
  • Safe-harbour trap: Google India v Visaka Industries (2020) requires “actual knowledge” through a court order — not a private complaint — before intermediary liability kicks in. A common CLAT trap.
  • FCU cautionary tale: Kunal Kamra v UoI (Bombay HC, 2024) struck down the IT Rules’ Fact Check Unit amendment — a reminder that not every government “content regulation” passes constitutional muster.

Key Cases at a Glance

Case Proposition
Shreya Singhal v UoI (2015) Struck down § 66A IT Act; read down § 79 to require a court order or government notification for takedown.
Justice K.S. Puttaswamy v UoI (2017) Privacy is a fundamental right under Articles 14, 19 and 21 — foundational for regulating AI content.
Anuradha Bhasin v UoI (2020) Proportionality test; internet access intrinsic to Art 19(1)(a) and 19(1)(g).
Google India v Visaka Industries (2020) Intermediary protected until “actual knowledge” via court order.
Kunal Kamra v UoI (Bombay HC, 2024) IT Rules FCU amendment struck down as unconstitutional.
Kaushal Kishor v State of UP (2023) Grounds in Art 19(2) are exhaustive; no free-standing “additional” restriction on free speech.

Mnemonic — "LABEL"

Leo (Puttaswamy) · Art 19(2) grounds are closed · Bombay HC FCU strike-down (Kunal Kamra) · Exact knowledge via court order (Visaka) · Lesser restrictive means — disclosure > prior restraint.

Closing date for the SGI consultation: 7 May 2026.

The Principle of Proportionality — Applied

If the CLAT 2027 passage runs a principle such as: “A regulation restricting speech must be narrowly tailored to a legitimate aim and satisfy the proportionality test”, and the fact pattern gives you MeitY’s SGI-labelling rule — the answer should be: likely constitutional. The rule regulates (not prohibits) satire and political commentary, requires disclosure rather than deletion, and serves the legitimate aim of preventing deception. Compare this with a hypothetical “ban on all AI-generated political speech” — that would almost certainly fall.

Test Yourself: 10-Question CLAT Practice Quiz

Ten CLAT-style questions on this story — factual recall, constitutional articles, and principle-application reasoning. Tap an option to lock it in; explanations appear on submit.

Practice Quiz — 10 CLAT-Style Questions

Click an option to reveal the answer and explanation.

Source: MediaNama · MeitY Press Release · LiveLaw · Khaitan & Co Analysis.

Quick answers

The 'safe harbour' provision that shields online intermediaries from liability for third-party content is located in which Act?
IT Act, 2000 (Section 79). Section 79 of the IT Act, 2000 grants conditional safe harbour to intermediaries that follow due-diligence norms.
In the landmark Shreya Singhal v Union of India (2015) ruling, the Supreme Court struck down:
Section 66A of the IT Act. Section 66A, which criminalised 'offensive' online messages, was struck down as violating Article 19(1)(a).
The new MeitY amendments deal principally with 'SGI' — Synthetically Generated Information. Which problem do SGI rules primarily target?
Deepfakes and AI-generated disinformation. SGI labelling rules are aimed at making deepfakes and AI-generated content identifiable.
Under the draft amendments, AI labels on SGI must be 'continuously and clearly visible'. The October draft had stipulated that labels should cover at least what share of a visual display or audio duration?
10%. The October draft specified a 10% minimum threshold; the final notification replaced this with a qualitative 'prominent and perceivable' test.
In Justice K.S. Puttaswamy v Union of India (2017), the Supreme Court held that the right to privacy — a basis for regulating AI content — is a fundamental right under:
Articles 14, 19 and 21. The 9-judge bench recognised privacy as intrinsic to Articles 14, 19 and 21.
Article 19(1)(a) — freedom of speech and expression — can be restricted only on the grounds listed in Article 19(2). Which of the following is NOT a ground under Article 19(2)?
Economic interest of the State. Art 19(2) grounds are: sovereignty/integrity, State security, friendly relations, public order, decency/morality, contempt of court, defamation, incitement to offence.
Share this article
CLAT Gurukul
Written by CLAT Gurukul

Ready to Crack CLAT?

This article covers just one topic. Our courses cover the entire CLAT syllabus with 500+ hours of live classes, 10,000+ practice questions, and personal mentorship from top faculty.

500+Hours of Classes
10,000+Practice Questions
50+Mock Tests
Start your CLAT prep with a free 5-day demo course Start Free Trial →