India’s latest amendment to the Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules have created quite a stir for significantly tightening platform obligations.
Intermediaries must now remove or disable access to unlawful content within three hours of receiving either a court order or a “reasoned intimation” from an authorised government officer. Certain urgent complaints—particularly involving impersonation, nudity, morphed content, and non-consensual sexual content—now have to be fast-tracked, with the government demanding that social media platforms pull down such content within two hours.
Other user grievances must be acknowledged within 24 hours and resolved within seven days.
The amendment also introduces new obligations regarding synthetically generated information (SGI), requiring platforms to implement technical and organisational measures to prevent the creation or dissemination of unlawful deepfakes. It also mandates labelling of permissible AI-generated content, subject to limited exceptions for routine edits and minor enhancements.
The government has framed these changes as necessary to combat deepfakes, impersonation, and cybercrime. However, critics argue that the amendment revives an unresolved controversy: using Section 79 of the IT Act as a ransom to force Big Tech’s compliance.
Reviving the Old Debate
Section 79 is essentially a safe-harbour provision. Originally enacted to protect platforms from liability for user-generated content, Section 79 grants immunity so long as intermediaries exercise due diligence and act “expeditiously” upon receiving actual knowledge of unlawful content through a court order or government notification.
If they fail to comply, they risk losing safe-harbour protection and may be exposed to legal liability under the IT Act and other laws.
Pranesh Prakash, an independent tech policy advisor and an affiliated fellow at Yale Law School, tells AIM that this liability framework is now functioning as a coercive tool rather than a shield.
“Section 79 of the Information Technology Act does not provide the government powers to block anything,” Prakash explains. “There’s Section 69A of the IT Act, which provides such powers.”
Section 69A is India’s formal blocking provision. It empowers the central government to block public access to information on grounds such as sovereignty, public order, and national security. Crucially, it operates under the Blocking Rules, 2009, which require written reasons, committee review, and procedural safeguards.
Unlike Secrion 66A, which the Supreme Court struck down in 2015 for its vagueness of criminalising any “offensive” electronic message, Section 69A has procedural safeguards subject to judicial review.
Prakash contends that the canon of “reasoned intimation” from an authorised officer under amended Rule 3(1)(d) of the IT Rules effectively allows the state to block access without invoking Section 69A’s procedural safeguards. As the platforms risk losing safe-harbour protection in case of non-compliance, they have little choice but to remove content even when the notice is issued by an executive officer rather than a court.
Supratim Chakraborty, Partner at Khaitan & Co, also flags interpretive uncertainties. “Much will depend on how regulators interpret applicability in practice and how platforms operationalise compliance without drifting into defensive over-moderation,” he says.
The shortened compliance window further makes seeking legal clarity or challenging questionable unviable.
Moreover, a deeper concern is transparency: notices issued under Section 79 are visible only to platforms or intermediaries, not to the content's original poster.
“What the government is asking social media intermediaries to remove is not public information,” Prakash explains. “It’s not communicated to the people whose content is being removed. So it’s a secret order….without any contestation by the parties involved.”
“How can the government engage in censorship without telling us what it is censoring?” he wonders.
Another structural question is who determines whether content is unlawful. For instance, even police officers (not below the rank of Deputy Inspector General) may issue such notices if specially authorised, and both central and state governments may designate officers or agencies for this purpose.
How Section 79 is Used in Practice
In March 2025, X Corp (formerly Twitter) filed a writ petition in the Karnataka High Court challenging the use of Section 79(3)(b) of the IT Act. X argued that the notice-based regime was effectively a parallel blocking mechanism without Section 69A’s procedural safeguards, allowing executive officers to function as de facto arbiters of what content is lawful. It flagged that the officer issuing content takedown notices effectively becomes the judge.
However, in September 2025, the Karnataka High Court dismissed X Corp’s petition. The court held that the government’s mechanism for issuing takedown notices falls within its statutory authority and does not violate the law. It also stated that regulating social media is legitimate and that platforms must comply with lawful takedown requests.
While the rules may be drafted due to defamation concerns, their real-world implications are far and wide.
Last March, the Ministry of Railways issued at least seven takedown notices to platforms, including X, YouTube/Google, and Facebook/Instagram, demanding removal of train-related videos showing crowding, vandalism and other incidents.
Three of these notices have become part of X’s legal challenge.
Similarly, in February last year, the Ministry of Railways ordered X to remove 285 links containing videos showing casualties from the New Delhi Railway Station stampede.
Critics argue that these examples demonstrate how the intermediary liability route is being used to effect removals without the safeguards built into Section 69A.
Implementation Challenges
Alongside constitutional concerns, implementation challenges remain significant.
The amendment requires platforms to deploy “reasonable and appropriate technical measures” to prevent unlawful synthetic content and to embed metadata and labelling tools for compliant AI content. That’s a tough ask given that AI detection technologies remain imperfect, particularly when content is modified or reprocessed.
Chakraborty also notes that the operational implications are substantial. “Operational implications are significant, particularly given the short transition window. Platforms will need to embed metadata and labelling tools, deploy automated detection systems, and recalibrate grievance and takedown workflows within compressed timelines.”
The amendment certainly has been designed to tackle AI pitfalls, especially in the wake of xAI’s Grok model publishing inappropriate and obscene images of users on social media feeds. According to a 2024 McAfee survey, 75% of Indians online have seen some for of deepfake content.
But in doing so, it has reopened a foundational dispute about due process, transparency, and the limits of executive power in regulating online speech.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.