‘The Sufferer Is The Individual’: Former MeitY Official On India’s Fast Takedown Rules

India’s internet takedown system is becoming faster, but the question is whether it is becoming too easy to use.

With the government’s takedown deadline now reduced to three hours, and the Sahyog portal being integrated with platforms such as Meta through an API, content can potentially move from a government request to removal at unprecedented speed. But what happens when there is little or no human intervention in that process?

Rakesh Maheshwari, a former senior official at the Ministry of Electronics and Information Technology (MeitY) who was involved in India’s information technology policy framework, says platforms need to retain a “human in the loop”.

If requests are simply pulled through an API and automatically acted upon, he warns, governments could be encouraged to make more requests because they know they will not be challenged. “The sufferer is the individual,” he says, arguing that such a system could work against freedom of speech and expression.

Maheshwari draws a distinction between India’s two key takedown mechanisms. Section 69A of the Information Technology Act provides a centralised blocking process with a five-member committee, notice and representation before a final decision. Section 79(3)(b) operates more broadly and in a decentralised manner, allowing authorised government agencies and police to seek the removal of unlawful content.

That distinction becomes particularly important when takedowns involve protests, journalism or content where the question of what is “lawful” is less straightforward. Maheshwari acknowledges that governments, journalists and platforms can all get things wrong. But with viral content, waiting days for fact-checking may mean the damage has already been done.

His answer is not to slow the system down entirely, but to build in checks, transparency and procedural safeguards — including verifying whether a request came from the right person, cited the right law and followed the required process.

In this conversation with Decode, Maheshwari explains how India’s blocking framework evolved, what the three-hour takedown window means for platforms, and why greater transparency is necessary when content — including journalistic and protest-related material — is removed.

Here are the edited excerpts from the interview.

Could you take us back to the early days of India’s content-blocking framework? What problem was Section 69A originally designed to solve, and how was the government expecting it to be used?

Section 69A and the 2009 Blocking Rules were designed at a time when most of the content was on websites rather than apps. The grounds for blocking were also much narrower than those under Article 19(2), which has a broader societal and country-level impact.

The Ministry of Electronics and Information Technology (MeitY) was given the overall responsibility because it was considered a relatively neutral ministry. A five-member committee was created, including a technical representative, with members at the Joint Secretary level to bring seniority, authority, awareness and accountability to the process.

A defined mechanism was also put in place for blocking requests, with a specific form and nodal officers across states and key ministries. These officers were required to obtain approval within their respective offices before making a request.

The process involved three stakeholders: the requesting agency, the platform hosting the content and the five-member committee. A notice was issued to the person who posted the content or, if they could not be identified, to the platform. The committee would hear the request and the platform's representation before deciding whether the content attracted Section 69A and should be blocked.

The committee's recommendation then went to the MeitY Secretary for approval, after which the content could be removed, an app could be taken down from an app store, or Internet Service Providers (ISPs) could be directed to block access.

In emergencies, the matter could go directly to the Secretary, followed by a post-facto committee meeting within 48 hours. In principle, this provided reasonably good checks and balances.

Section 69A is a structured, centralised blocking mechanism. Section 79(3)(b), on the other hand, operates more like a distributed takedown mechanism. Why were these two mechanisms needed, and what is the important difference between them?

Both mechanisms existed from the beginning, but their purposes and processes are different.

Originally, individuals could bring content to a platform’s notice, with takedowns linked to the broader grounds under Article 19(2), including obscenity, defamation and contempt of court — grounds not covered under Section 69A. The Supreme Court’s Shreya Singhal judgment changed this, limiting such requests to authorised government agencies or courts.

Section 69A is a formal order backed by notice, platform representation, committee consideration and approval — essentially a collective, quasi-judicial process. Section 79(3)(b), meanwhile, deals with a much wider range of unlawful online content, including material prohibited under laws originally written for the physical world.

Because of this, takedown requests should come from authorised officers of the relevant government department, ministry or police, rather than just anyone within an agency. With many government agencies and platforms now involved, MeitY also took on a coordinating role — providing clarity on designated contact persons and, ideally, a dashboard showing how many requests were made, rejected and resulted in content being removed.

Recent reporting has highlighted an API integration between the Sahyog portal and Meta, which can make the takedown process much more automated. When the deadline for action has been reduced from 36 hours to three hours, what safeguards do you think are necessary to prevent automated takedowns from becoming excessive?

I would not say that human review has stopped. Platforms such as Meta are still supposed to have human review. But the three-hour timeline creates a real challenge, even as the urgency caused by viral content has to be recognised.

Platforms need a human in the loop. Algorithms can help assess probabilities and identify doubtful or borderline cases where human intervention is needed. Large platforms such as Meta may be able to handle frequent requests, but smaller platforms may struggle. There also needs to be a mechanism confirming that a request has actually been received.

If a platform simply uses an API to pull requests and automatically act on them, there is a concern that governments may be encouraged to make more requests because they know they will not be challenged. Once such a tool is in place, it can work 100% of the time as long as the URL is valid. That can go against freedom of speech and expression, with the individual ultimately bearing the consequences.

Not every government request will be a borderline case, and a platform may believe the government has done its due diligence. One option could be temporary removal followed by offline review, with the content restored if it is found not to warrant removal.

Even if platforms cannot judge the substance of every government request, they can still check the process: Is it from the right person? Is it in the right format? Does it cite the right law? Why is the content considered unlawful?

Right person, right format, right law and a clear reason for illegality — these procedural checks can act as important safeguards.

We have seen takedowns of protest-related content and also journalistic reports. When such content is removed under a three-hour process, does this create a worrying precedent for journalism and freedom of expression?

I wouldn’t necessarily call it a scary precedent if the process is handled maturely by both sides.

With journalism, time is of the essence. An old photograph or video can be given a new context, making it difficult to determine who is right at face value. Journalists and newspapers generally do their jobs responsibly, but governments can be wrong, journalists can be wrong, and platforms can also be wrong.

There is no easy solution. If you wait three days for fact-checking, the damage from viral content may already be done and cannot necessarily be undone. At the same time, when you look at an issue from a law-and-order perspective, you may naturally favour the government's position to some extent. But excessive action is still wrong.

The answer is checks and balances and greater transparency. We need to know how many accounts or pieces of content were taken down, on what grounds, who initiated the requests, whether the person making the request was authorised, and whether the required Secretary-level reviews are actually happening.

Section 69A has stronger checks and balances. Section 79 is more decentralised, with state police also able to make requests. Police can be very effective in cases of individual harm, but when it becomes a question of societal harm, the government's role becomes more significant, and political or establishment considerations can enter.

Section 79's process is fairly robust, but it is not always followed in its full spirit, and there isn't enough transparency. We don't need details of every individual post, but we do need statistical clarity on how the system is being used and whether authorised officials are actually the people authorised to make such decisions in the physical world.

These checks may get you to 80% accuracy, but you will never get to 100%.

Once content has been taken down, what should transparency and the right to challenge look like for the affected user?

The affected user should be informed, but they don't necessarily need to receive everything contained in the government's request. There is a trusted or fiduciary kind of relationship between a government or law-enforcement agency and a platform. If a government agency makes a request to a platform, the platform should not necessarily pass the entire request verbatim to the user.

The user needs enough information to understand where to go if they have a grievance. They should be able to approach the platform, its grievance officer, the Grievance Appellate Committee, or the Indian courts.

At a minimum, the user should know that the content was removed, who initiated the request and, where appropriate, why it was considered unlawful. That gives the person enough information to contest the action.

The problem is that platforms are inconsistent — sometimes sharing everything with users and sometimes nothing. Platforms should therefore have a clear, consistent policy on what and how they inform users, or the government should specify the minimum information that must be communicated.