For years, the debate over internet censorship in India has revolved around a familiar question: does the government have the power to take down online content? The answer, in certain circumstances, is clearly yes. Indian law provides the State with mechanisms to restrict access to unlawful or harmful online material, subject to statutory requirements and constitutional limitations. But that is no longer the most important question. The more urgent question is what happens when that power is exercised at extraordinary speed, at enormous scale, with limited transparency and with increasingly automated compliance by private platforms.
Recent figures reported through an RTI investigation by The Indian Express suggest that India is witnessing an unprecedented expansion of government-backed content restrictions. Between March and July 2026, government agencies issued nearly 1.95 lakh blocking directions to Instagram, Facebook and YouTube. According to the report, this amounted to roughly 1,275 directions every day, or approximately one direction every 68 seconds. Instagram alone received nearly one lakh directions, Facebook around 80,000 and YouTube approximately 15,000.
The scale becomes even more striking when compared with the previous period for which comparable figures are available. RTI data accessed by The Indian Express showed that between October 2024 and October 2025, 2,312 blocking orders were sent to 19 online platforms through the government’s Sahyog portal. That amounted to roughly six orders a day. In the following five months, directions to just three major platforms rose to almost 1.95 lakh. Such a dramatic increase cannot be treated merely as an administrative statistic. It raises a much larger question about the changing relationship between the State, technology companies and citizens in India’s digital public sphere. And the timing makes that question even more difficult to ignore.
When protest moves online, censorship follows
A significant proportion of the latest blocking directions reportedly coincided with the student protests over examination irregularities and leaks in Delhi. The Indian Express, citing a senior government official, reported that a considerable share of the directions were issued as the NEET protests gathered momentum, particularly on Instagram. This is important because Instagram was not simply being used as a space in which the protests were discussed after they happened. It became one of the principal spaces through which the movement itself was organised.
The material compiled in the research shows how the Cockroach Janta Party’s social-media presence became central to the mobilisation, with protest announcements, videos, memes, explainers and other political content circulating rapidly among young users. The account reportedly grew to millions of followers and played an important role in mobilising people for demonstrations, including the July 20 “Sansad Chalo” march. That transformation of social media into a political organising space has a profound constitutional consequence.
A protest today does not begin only when people gather on a street. It begins when people communicate. They decide where to meet, circulate information about police action, share videos from the ground, publicise their demands and tell others what is happening. The digital space has therefore become part of the infrastructure of democratic protest itself. Consequently, restricting that space can affect far more than an individual’s ability to upload a post. It can affect the ability of a movement to organise, document itself and communicate with the wider public. This is why the reported restrictions during the protests deserve particularly close scrutiny.
Article 14, for instance, reported that journalist Sadhika Tiwari’s video showing a student allegedly being beaten by police was blocked on X following a government order. The video, according to the report, showed a crying student and did not contain hate speech, misinformation or graphic imagery. Around the same period, Meta temporarily blocked the Instagram account associated with CJP, which reportedly had more than 23 million followers, although the account was subsequently restored after criticism. The significance of such incidents lies not merely in whether one particular video was lawfully restricted. It lies in the fact that videos documenting State action during a protest are themselves part of the democratic record.
If citizens cannot see what happened on the streets, they cannot independently evaluate competing claims about what happened there. If videos documenting police conduct disappear before they can circulate widely, public debate becomes dependent increasingly upon official accounts of events. That is precisely where the problem of censorship becomes a problem of democracy.
The three-hour problem
The most troubling development, however, may not be the number of takedown directions alone. It is the shrinking amount of time available to platforms to respond to them. As Article 14 reported, the period within which intermediaries are expected to comply with certain government requests has been progressively reduced, eventually reaching three hours in 2026. The investigation linked this shrinking compliance period with the sharp increase in government takedown directions. Three hours may sound like an administrative efficiency measure. It is not difficult to see why it becomes something much more consequential when the subject is constitutionally protected speech.
Imagine a government order directing a platform to remove a journalist’s report. Before complying, the platform would ideally need to determine whether the issuing authority has jurisdiction, whether the order identifies the material adequately, whether the statutory requirements have been met, whether the content actually falls within the legal grounds for restriction and whether the restriction is proportionate. The affected journalist may also need an opportunity to respond. A court may eventually have to examine the order. Yet the platform is given only a few hours to act.
Under those circumstances, the rational institutional response is unlikely to be a careful legal assessment. It is likely to be compliance first and review later. That creates an important constitutional problem. A system that imposes severe consequences for delayed compliance but relatively weak consequences for wrongful removal creates an obvious incentive for platforms to err on the side of deletion. The consequence is what free-speech activists have long warned about: over-removal.
Lawful speech does not have to be formally prohibited for censorship to occur. It is enough for platforms to become sufficiently afraid of retaining controversial material that they begin removing content whenever there is doubt.
Sahyog and the transformation of censorship into infrastructure
The government’s Sahyog portal illustrates another aspect of this changing architecture. Introduced in October 2024, Sahyog was designed to streamline communication between government authorities and online intermediaries. In itself, there is nothing inherently problematic about creating an administrative mechanism through which lawful government directions can be communicated to technology companies. But administrative efficiency cannot substitute for constitutional safeguards.
Article 14 reported that X challenged Sahyog before the Karnataka High Court, arguing that the system effectively facilitated censorship by allowing government requests to reach platforms through a centralised mechanism. The Karnataka High Court rejected the challenge in July 2025 and characterised Sahyog as a facilitation mechanism rather than an independent censorship authority. That judicial characterisation is important, but it does not end the constitutional debate.
A communication system may technically be nothing more than a channel. The real constitutional question concerns what happens after the communication reaches the platform. If a government direction arrives through an official portal, carries the appearance of State authority and has to be acted upon within three hours, the practical ability of a platform to question that direction becomes extremely limited. The portal may facilitate communication.
But when that communication is backed by an extremely short compliance window and automated enforcement, the overall system can begin functioning as something considerably more powerful: an infrastructure through which State decisions about speech are translated rapidly into restrictions on citizens’ access to information.
When the algorithm becomes the censor
This is perhaps the most disturbing development. According to The Indian Express, Meta integrated its systems with Sahyog to comply with the three-hour requirement, allowing content identified through government directions to be restricted or removed automatically rather than waiting for a separate human review before compliance. Automation is attractive because it makes compliance efficient. But constitutional rights are precisely the kind of interests that should make us suspicious of efficiency without accountability.
A human decision-maker can, at least theoretically, ask whether a government direction is valid. An automated system cannot meaningfully interrogate the constitutional legitimacy of the State’s demand. It receives an instruction and executes it. The danger is therefore not simply that the government can order content to be removed. The danger is that a government instruction can increasingly be converted into an automated technical command. The chain becomes remarkably simple: the State issues the direction, the platform’s system receives it, the content is restricted and the user discovers that the speech has disappeared.
A citizen cannot challenge a decision that they cannot see
The question of notice is therefore central. Alt News has documented instances in which users received generic notifications stating that content had been restricted pursuant to a legal request or through an automated process, without being provided meaningful information about the specific legal basis or authority involved.
This creates a basic procedural problem. If a person does not know who ordered the restriction, they cannot determine whether the authority had jurisdiction. If they do not know which law was invoked, they cannot assess whether the restriction is legally sustainable. If they do not know why their particular content was unlawful, they cannot meaningfully challenge the decision. And if they cannot access the order itself, they may not even know what exactly they are supposed to challenge. This is where transparency stops being a policy preference and becomes a constitutional necessity. The State cannot reasonably expect citizens to defend their right to speak against an order whose existence, reasoning and legal basis remain hidden from them.
The disappearance of journalism is more serious than the disappearance of content
The consequences become particularly serious when the material being restricted is journalism. Article 14 reported that in March 2026 alone, Facebook pages belonging to at least six news and digital publications were blocked. The report also identified restrictions involving Greater Kashmir, Rising Kashmir, Kashmir Life, Molitics India, National Dastak and 4PM News Network, among others. The Wire, Maktoob Media and Peek TV have similarly reported restrictions affecting their content.
The problem here is fundamentally different from the removal of ordinary user-generated material. A journalist reporting on a protest is not simply generating “content”. They are creating a record of a public event. A report about Kashmir may become part of the historical record of the region, a video documenting police action may later become evidence in litigation, or a report concerning government policy may inform electoral choices. When such material disappears without clear explanation, the damage therefore extends beyond the individual publisher.
The public loses access to information that may be essential to holding power accountable. That is why the freedom of the press cannot be protected merely by allowing newspapers to continue printing. In 2026, press freedom also requires journalists to be able to reach audiences through the digital platforms on which those audiences increasingly depend.
The State’s concern about decentralised communication is not irrelevant — but it is not enough
The controversy surrounding Bitchat, Briar and Bridgefy demonstrates that the problem extends beyond mainstream social-media platforms. During the Jantar Mantar protests, demonstrators reportedly turned to Bluetooth-based mesh communication applications that could facilitate communication between nearby devices without relying entirely on conventional mobile internet infrastructure.
The government’s concern was that such systems could make lawful interception, identification and investigation more difficult. According to reports, the Indian Cyber Crime Coordination Centre issued directions seeking the removal of Bitchat, Briar and Bridgefy from major app stores and also sought action against Bitchat’s GitHub repository. Medianama reported that the notices provided intermediaries with three hours to comply, although officials later orally instructed companies not to enforce the directions.
ThePrint reported that the I4C notice to GitHub invoked Section 79(3)(b) of the Information Technology Act read with Rule 3(1)(d) of the IT Rules and sought removal of Bitchat’s repositories within three hours. There is a legitimate public-interest argument here. Communication technologies can be abused for criminal activity, and the State cannot be expected to remain completely powerless when technology is used to facilitate serious offences. But that cannot be the end of the analysis.
The fact that a technology can be misused does not establish that the technology itself should be suppressed. A telephone can be used to organise a crime. A messaging application can be used to coordinate violence. Social media can spread misinformation. None of these facts, by themselves, justify eliminating the technology. The State must still establish why the restriction is necessary, why less restrictive alternatives would not work and why the particular application or communication mechanism presents a sufficiently serious and demonstrable threat. Otherwise, “public order” becomes a justification capable of swallowing the right to communicate altogether.
The constitutional framework is not optional
India does not indulge this issue without constitutional guidance. Article 19(1)(a) guarantees freedom of speech and expression, while Article 19(2) permits reasonable restrictions on specified grounds. The Supreme Court’s decision in Shreya Singhal v. Union of India remains a foundational judgment in understanding the limits of online speech regulation.
The Court’s decision is important precisely because it recognises that vague or overbroad restrictions can have a chilling effect on legitimate expression. The existence of a statutory power, therefore, cannot by itself answer the constitutional question. The State must still demonstrate that the restriction falls within the law and survives constitutional scrutiny. That is particularly important where the restriction affects political expression.
Political speech lies at the heart of democratic participation. Criticism of government policy, reporting on protests, documenting police action and exposing administrative failures cannot be placed in the same category as genuinely unlawful material merely because they are inconvenient or controversial. The State must be able to distinguish speech that is unlawful from speech that is merely uncomfortable. That distinction is the foundation of a functioning democracy.
Another concern emerging from the reporting is the increasingly routine character of mechanisms originally justified by urgency. Emergency powers exist for a reason. There are situations in which authorities may need to act quickly to prevent serious and immediate harm. But emergency mechanisms become constitutionally dangerous when the exception starts functioning as the ordinary route for restricting speech.
Article 14 reported concerns among technology-policy experts that emergency blocking mechanisms could increasingly become part of routine online governance, particularly because affected parties may not receive the underlying orders and therefore have limited opportunities to challenge them. The constitutional problem can be stated simply. An emergency procedure should make the State faster when speed is genuinely necessary. It should not make the State less accountable because it has chosen to act quickly. If anything, extraordinary powers should trigger stronger review precisely because they reduce the opportunity for the affected person to respond beforehand.
Platforms are not innocent bystanders
The responsibility, however, does not rest with the government alone. Technology companies have enormous power over what citizens can see, publish and circulate. They therefore have their own obligations of transparency. The Indian Express reporting indicates that Meta generally provides users in several jurisdictions with information when governments request restrictions on their content, but that such disclosures are not provided in India and some other markets because of legal and regulatory considerations. That should concern anyone who believes that users have a right to understand why their speech has been restricted.
A platform should not be able to tell a journalist merely that their report is unavailable because of a “legal requirement”. The journalist should know whether the government ordered the restriction, which authority issued the order and which law was invoked. Transparency does not necessarily prevent compliance. A platform can comply with a lawful government direction while still informing the affected user that the government sought the restriction. Indeed, transparency may be one of the few mechanisms capable of ensuring that government censorship does not become invisible.
The reported proposal by X to make government requests more visible to users therefore deserves attention. According to recent reports, Elon Musk announced that X would begin making government requests to restrict or remove content more transparent, including information about the requesting authority and, where available, the legal basis for the request.
The Indian government has understandably maintained that platforms operating in India remain subject to Indian law. But that is not inconsistent with transparency. The two principles can coexist. If the State has lawfully ordered a restriction, there should be no reason for the process to remain invisible. Government censorship should not become anonymous censorship.
India must decide what kind of digital democracy it wants
There is a legitimate State interest in regulating harmful online material. There is a legitimate public interest in combating misinformation, incitement, deepfakes, threats and criminal coordination. Technology companies cannot claim absolute immunity from the laws of the countries in which they operate. But none of these propositions answers the central constitutional question.
How much power should the State possess to determine what citizens are allowed to see, publish and circulate — and what safeguards must accompany that power?
Nearly 1.95 lakh blocking directions in five months, as reported by The Indian Express, should force that question into the centre of India’s democratic debate. The answer cannot simply be that the government has followed a procedure. A procedure that is too fast to meaningfully challenge, too opaque to independently scrutinise and too automated to permit genuine review can satisfy the appearance of legality while undermining the substance of constitutional protection.
Indian law therefore needs stronger safeguards around government-directed content restrictions. Orders should clearly identify the legal provision relied upon, the authority issuing them and the reasons for restricting the particular material. Affected users should ordinarily receive notice and have access to a meaningful appeal mechanism. Emergency orders should be subject to prompt independent review and should automatically lapse unless their continuation is justified. Platforms should publish greater information about government requests, including anonymised data that allows the public to assess how frequently such powers are being used and how often they are challenged or withdrawn. Most importantly, the burden cannot silently shift from the State having to justify censorship to the citizen having to prove why their speech should remain online.
The extraordinary rise in takedown directions, the targeting of protest-related content, the restrictions imposed on journalists and news organisations, the attempt to interfere with decentralised communication tools and the increasing use of automated compliance together point towards a digital environment in which the speed of censorship may be outpacing the speed of accountability.
Related:
JNU Cancels Venue, Not the Conversation: Umar Khalid’s book discussion moves outdoors amid protests
The Supreme Court’s Uneasy Middle Path on Protest Policing
Ken-Betwa protest exposes the cost of state-led displacement