UPSC Darpan

Science & TechnologyGS39 October 2026

MeitY Asks Platforms to Curb Manipulated Content as Vaishnaw Promises an AI Regulation Paper

Open in the app — quiz, notes, Mistake Vault हिंदी में पढ़ें

The news

New Delhi. The Ministry of Electronics and Information Technology (MeitY) on Thursday, October 8, issued an advisory urging social media platforms to curb “manipulated or out-of-context content”, beyond the AI deepfakes already liable to takedown under earlier amendments to the IT Rules, 2021, The Hindu reports. Such content, it said, can “cause public confusion… disturb public order”. It precedes the Cockroach Janta Party’s October 10 protest in Delhi. The same day, IT Minister Ashwini Vaishnaw said a consultation paper on AI regulation would come within a month, centred on safety, deepfakes (a “nuisance”), a human-first approach and skilling. AI cannot be regulated “by law alone”, he said; India needs a “techno-legal solution”, legal duties backed by tools such as labels and detection.

The chain in one line: Generative AI makes realistic fakes cheap → IT Rules amended to require labels and fast takedown → MeitY widens the net by advisory before a protest → an AI regulation paper is promised

Static syllabus linkage

  1. Section 79 safe harbour depends on due diligence. Section 79 of the IT Act, 2000 shields an intermediary from liability for users’ content if it follows due diligence, and Section 79(3)(b) removes the shield if it ignores “actual knowledge” of unlawful content. The IT (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021 define that due diligence; Rule 3(1)(b) requires reasonable efforts to stop users sharing listed unlawful or false content.
  2. Shreya Singhal (2015) limits takedown. The Supreme Court struck down Section 66A of the IT Act as vague and read down Section 79(3)(b): “actual knowledge” means a court order or government notification, not a private complaint. It upheld Section 69A, the Centre’s blocking power, which needs recorded reasons.

Why UPSC loves this

  1. GS3 links IT with security threats through social media. The syllabus names “awareness in the fields of IT” and the “role of media and social networking sites” in internal security.

Prelims nuggets

  • Section 79 of the IT Act, 2000 gives intermediaries “safe harbour” from liability for third-party content, subject to due diligence.
  • In Shreya Singhal v. Union of India (2015), the Supreme Court struck down Section 66A of the IT Act and upheld Section 69A.
  • Amendments to the IT Rules, 2021 notified in February 2026 define “synthetically generated information” and require it to be labelled.
  • Article 19(2) includes public order among grounds for restricting speech; misinformation is not a listed ground.

Analysis

  1. “Out-of-context” content is harder to police than a deepfake. A deepfake is a fabricated file, so it can be defined, labelled and detected. A real video with a false caption is not; its harm lies in the claim attached. Judging context makes platforms arbiters of truth, so this advisory will be applied unevenly.
  2. Lens — Innovation and safeguards: fight fakes through law, not advisories. The harm is real: a fake released before a tense protest can spark violence. But an advisory issued on the eve of an opposition-linked protest invites the charge that “public order” is cover for limiting political speech; the Bombay High Court struck down the government fact-check unit in 2024 on such worries. A thoughtful officer would write narrow definitions, reasoned orders and an appeal into the new framework.
  3. Placing the burden on industry has a price. The Minister said “the bulk of the responsibility” lies with industry. Big platforms can afford detection and provenance tools; Indian start-ups may not, so heavy compliance favours incumbents. Duties should scale with risk and size, as in the EU AI Act.

Possible Mains question

Examine whether India’s intermediary-rules approach can deal with deepfakes, or whether a dedicated AI law is needed. (15 marks, 250 words)

Model approach

  1. Directive — Examine. Test both routes, then conclude.
  2. Introduction — the October 8 advisory and promised AI paper mark a shift towards law. Define a deepfake.
  3. Body — the IT Rules give speed through labels and takedown, but are delegated legislation. Value addition: Shreya Singhal (2015) on “actual knowledge”.
  4. Body — advisories on vague categories like “out-of-context” risk chilling political speech. Bombay High Court, 2024.
  5. Body — a risk-based law can share duties across the chain. Diagram: developer → platform → user, marking the duty at each stage (watermark, label, report).
  6. Conclusion — legislate narrowly, with reasoned orders and appeal. Parliament, not advisories.

Administrator's brainstorm

As a District Magistrate, a deepfake of a communal incident spreads on the eve of a protest. What do you do?

Verify fast through the cyber cell and post an official clarification on the same platforms. Send a reasoned takedown request through the nodal officer, recording the public-order threat. Avoid a blanket internet shutdown, and preserve evidence to prosecute the source.