MeitY Advisory on Manipulated Content: Platform Accountability and Free Speech

What the advisory seeks to change

According to the report, MeitY has urged platforms to exercise enhanced diligence towards content that is false, misleading, manipulated or deliberately presented out of context. Its focus extends beyond AI-generated deepfakes to old photographs and videos passed off as current events.

The advisory highlights possible public confusion and public-order consequences, and asks platforms to prioritise reports from authorised government agencies. The report describes earlier amendments as addressing AI-generated content and providing users with an option to label AI-generated imagery; the advisory’s broader language should not itself be treated as a new statutory amendment.

• Misleading context can arise without any AI-generated or digitally altered material.
• Prioritising a government complaint is analytically distinct from establishing that the reported content is unlawful.
• The protest-related timing makes safeguards for political criticism and peaceful mobilisation particularly important.

Advisory, blocking direction and safe-harbour notice are different

An executive advisory does not, merely by being issued, acquire the force of an Act or duly notified rule. It may communicate the government’s interpretation of existing duties, but any coercive consequence must be traceable to valid law; an advisory cannot independently create a new offence or override constitutional protections.

Section 69A concerns statutory blocking directions issued through a prescribed process. Section 79(3)(b), by contrast, concerns circumstances affecting an intermediary’s exemption from liability after legally relevant knowledge or notification. It is not the same legal instrument as a Section 69A blocking order.

The report describes government referrals under Section 79(3)(b) as not themselves binding removal orders, while noting automatic takedowns by Meta under compressed timelines. The essential qualification is that failure to act on a legally valid notification can affect safe-harbour protection; loss of that protection is not, by itself, a finding of liability for the underlying content.

• Examine the legal provision invoked, the issuing authority and the alleged underlying illegality.
• Distinguish a complaint, a statutory notification and a formal blocking direction.
• Do not infer universal removal obligations from the reported compliance window without examining the applicable rule.

Constitutional limits and judicial safeguards

Article 19(1)(a) protects expression, including criticism of public institutions. Public order is a permissible ground for restriction under Article 19(2), but neither falsity nor the use of AI is an independent ground listed there. Restrictions require legal authority and must satisfy constitutional standards of reasonableness and proportionality.

In Shreya Singhal v. Union of India, the Supreme Court read down the actual-knowledge standard under Section 79(3)(b) to knowledge through a court order or notification by the appropriate government or its agency, concerning unlawful acts relatable to Article 19(2). The judgment also upheld Section 69A and its blocking framework in light of their statutory grounds and safeguards.

References to public confusion or apprehension cannot automatically substitute for demonstrating the legally relevant harm. Satire, parody, documentary archives and legitimate political disagreement require careful contextual assessment rather than category-wide suppression.

• A public-order restriction requires a sufficiently proximate connection to the anticipated harm.
• Government-origin complaints must remain subject to legal scrutiny rather than a presumption of correctness.
• Notice, reasons and meaningful opportunities to challenge action are central due-process safeguards, subject to lawful exceptions.

Why intermediary incentives matter

The report states that a February amendment shortened compliance timelines to 1–3 hours and that Meta has automatically removed posts referred through the reported channel. Tight deadlines, uncertainty about illegality and fear of losing safe harbour can together incentivise over-removal.

At the same time, harmful impersonation and deceptive media can spread rapidly, making delayed intervention ineffective. Accountability therefore requires both rapid handling of clearly unlawful harm and safeguards against indiscriminate suppression.

A calibrated response can include contextual labels, provenance information and user warnings where appropriate, while complying with valid removal or blocking requirements. Automated detection should support, rather than replace, contextual legal assessment in disputed cases.

• Authenticity assessment asks whether material is genuine; legality assessment asks whether its dissemination violates law.
• Regional-language and cultural context are essential for distinguishing deception from satire.
• Transparent appeals and correction mechanisms can reduce the lasting effects of erroneous takedowns.