The Bombay High Court’s split verdict on government fact-checking

The first blog post in this series examined how the two Judges of the Bombay High Court understood the scope and effect of Rule 3(1)(b)(v) of the IT Rules in significantly different ways. Justice Patel characterized the Rule as a broad rule that could allow for the indirect censorship of any content concerning the Union Government by threatening to remove an intermediary’s safe harbour. In contrast, Justice Gokhale interpreted the Rule to apply to cases where users intentionally share misinformation and concluded that the Rule did not require intermediaries to remove the content flagged by the Government’s Fact Checking Unit (“FCU”).

As a reminder, Rule 3(1)(b)(v) caused intermediaries to lose their safe harbour if they did not make reasonable efforts to remove content that:

“deceives or misleads the addressee about the origin of the message or knowingly and intentionally communicates any misinformation or information which is patently false and untrue or misleading in nature or, in respect of any business of the Central Government, is identified as fake or false or misleading by such fact check unit of the Central Government. (emphasis supplied)”

Once they had characterised the Rule, each Judge had to decide three notable free speech issues. First, is “fake, false or misleading” speech (as identified by the FCU) a class of speech that can be restricted under Article 19(2). Second, is the Rule unconstitutionally vague? Third, does the Rule constitute a proportional restriction on free speech and offer sufficient procedural safeguards against abuse?

False speech under the Constitution

Article 19(1)(a) provides for the freedom of speech and Article 19(2) sets out an exhaustive list of reasons for restricting speech (¶101, 153 Patel J.). These are the security, unity, and integrity of India and its friendly relations with other States, public order, decency or morality, defamation, contempt of court, or incitement to an offence. Thus, any restriction on speech must bear a nexus to one of these State interests. Crucially, as Justice Patel notes, “public interest” (or falsehood) is not a ground to restrict free speech under Article 19(2) (¶135 Patel J.). He further observes that in certain situations, where there exists a meaningful nexus with a legitimate aim in Article 19(2), false information may be restricted. For example, he asks, “Where might a piece of ‘fake news’ calling for an insurrection or an incitement to communal or other violence fall? Conceivably, this could well be within the Article 19(2) limits of ‘public order’ (¶164 Patel J.). But in such a situation, speech would not be restricted solely because it is false, but because it threatens public order – a legitimate State interest under Article 19(2). However, Rule 3(1)(b)(v) of the IT Rules restricts speech solely on the grounds that it has been flagged as false by the FCU. Thus, as Justice Patel concludes, “The impugned Rule takes up falsity per se, and restricts content on that ground divorced from, and untraceable to, any specific part of Article 19(2). That would be impermissible. (¶164 Patel J.). It is impermissible to restrict speech solely on the ground that it is false.

This is also a classic overbreadth analysis. Justice Patel’s opinion recognises that Rule 3(1)(b)(v) may restrict speech can be validly curtailed under Article 19(2) (e.g., false information that amounts to incitement).

However, because the Rule goes beyond this to also potentially restrict lawful speech that has no nexus with a State interest under Article 19(2), it is an overbroad speech restriction. As Justice Patel notes, the Rule“necessarily and axiomatically makes suspect and subject to identification with no reference to cause, effect or Article 19(2)’s constraints, every single digital data chunk that relates to the business of the Central Government (¶157 Patel J.). 

Justice Gokhale’s approach to this question takes some breaking down. First, she observes that false speech would not be free speech under Article 19(1)(a) of the Constitution itself (¶¶41, 55 Gokhale J.). She finds that the right to free speech does not include the right to abuse one another, as this would undermine societal harmony and national security. This is problematic because nothing in the text of Article 19(1)(a), which simply provides for the “freedom of speech” suggests that any speech, let alone false speech, does not constitute free speech. It is not as though the Indian Constitution does not allow restrictions on speech, or indeed even false speech. For example, the Constitution itself identifies defamation (a type of false speech) as a ground to restrict speech. But this is provided for in Article 19(2). Therefore, Article 19(2) itself captures the balancing required between free speech and other societal interests such reputation or national security. Hence, it is submitted that the more convincing interpretation of Article 19(1)(a) and Article 19(2) is that all speech is first protected as free speech by Article 19(1)(a) and only restrictable on the grounds set out in Article 19(2). 

Justice Gokhale’s second answer to Article 19(2) is to note that the Supreme Court in Shreya Singhal v. Union of India held that an intermediary will only lose its safe harbour if a Court or government agency directs it to remove content, and that such a removal direction must relate to an Article 19(2) interest (¶20 Gokhale). This makes sense, because if the content removed exceeded a permissible restriction under Article 19(2), the Court or government agency would be removing lawful speech. Justice Gokhale then concludes that because Shreya Singhal held that an intermediary can only be required to remove content that is validly restrictable under Article 19(2), Rule 3(1)(b)(v), which requires intermediaries to take down FCU flagged content, is compliant with Article 19(2) (¶21 Gokhale J.).

With respect, this is entirely circular logic. Shreya Singhal held that a Court or the Government can only restrict content under the IT Rules when it is constitutionally permissible under Article 19(2). Rule 3(1)(b)(v) is a ground to restrict speech under the IT Rules, a ground that has been added to the Rules after the Shreya Singhal verdict. One cannot conclude that Rule 3(1)(b)(v) is constitutionally valid merely because Shreya Singhal said that all restrictions under the IT Rules must be valid. The court must test the validity of the Rule and provide an independent justification for a conclusion of validity that is not already part of its premise. If Justice Gokhale’s logic is applied, constitutional reasoning would be as follows: laws must comply with the Constitution, this is a law, therefore it is compliant with the Constitution. This is not a tenable approach and entirely evades the tests set out by the Supreme Court to judge the constitutionality of a free speech restriction.


On the question of whether Rule 3(1)(b)(v) was unconstitutionally vague, the two Judges’ divergent interpretations of the scope of the Rule come to the fore. Because Justice Gokhale construed the Rule as applying to only situations where users intentionally shared false content, she found that this narrowed the scope of the Rule to within judicially manageable standards (¶31 Gokhale J.). She further held that a breach of the Rule would only occur if the user shared content with a reckless disregard for the truth or “actual malice” (¶37 Gokhale J.). Finally, Justice Gokhale concluded if content, no matter how critical of the Government was based on some “existent and not fake or known to be misleading” information, it would not be hit by the Rule (¶44 Gokhale J.).

In contrast, Justice Patel highlighted the difficulties in ensuring that content concerning the government was not “fake or false or misleading”. He pointed out that information about the government, indeed even government data concerning the economy, poverty, or health, was often within the realm of subjective assessment (¶119 Patel J.) He then referred to the Indian Evidence Act, and in particular the definition of “Proved”. The provision states that:

 “A fact is said to be proved when, after considering the matters before it, the Court either believes it to exist, or considers its existence so probable that a prudent man ought, under the circumstances of the particular case, to act upon the supposition that it exists. (emphasis supplied)”

Justice Patel observed that the issues of fact, proof, belief, probability, prudence, and supposition, which are central to determining facts in law, was absent in Rule 3(1)(b)(v) (¶115 Patel J.). He therefore found the terms “fake or false or misleading” to be impermissibly vague because they offered no judicially manageable standards to determine restrict speech and conferred excessive discretion on the FCU.

Given that both Judges’ interpretation of the vagueness question was fundamentally shaped by their understanding of how Rule 3(1)(b)(v) operated (and that has already been discussed extensively in Part I of this blog series), we can leave this issue to rest here.

Proportionality and procedural safeguards

Both Judges referred to Supreme Court precedent and acknowledge that Rule 3(1)(b)(v) must satisfy the test of proportionality. The divergence in their proportionality analysis is effectively captured by the necessity limb. Justice Gokhale found that Rule 3(1)(b)(v) is necessary because the Government is concerned about the negative impact of false information spreading on social media, which may threaten public order (¶54 Gokhale J.) However, this is not what the test of necessity requires. The necessity limb of the proportionality test requires the court to examine whether there exist any alternative measures that are equality effective. In other words, could the Government have adopted a less rights-restrictive but equally effective measure? For example, Justice Patel highlights that the Government, through the Press Information Bureau, is well equipped to independently clarify or rebut incorrect information concerning its workings (and in fact regularly does so) (¶182 Patel J.). Therefore, in his eyes, Rule 3(1)(b)(v) failed the test of necessity.

The last limb of the proportionality test concerns the existence of procedural safeguards. Justice Gokhale highlights that users whose content is taken down can approach the intermediary with a grievance, and the intermediary is required to address this grievance within 15 days (¶19 Gokhale J.). Further, she highlights that users can also approach the Government’s own Grievance Appellate Committee specifically tasked to deal with online content (¶29 Gokhale J.). Lastly, Justice Gokhale finds that merely because the FCU is staffed by members of the executive, and the information they are adjudicating the correctness of information that concerns the Union Government, it cannot be presumed that the Government will be biased (¶25 Gokhale J.). To substantiate this finding, Justice Gokhale refers Crawford Belly v. Union where the Supreme Court held that the Government’s appointment of an Estate Offer was not a ground to presume bias in the actions of the Officer (¶26 Gokhale J.).

Two points may be made here. First, the existence of procedural safeguards cannot be completely satisfied by the existence of an appeal process. There must exist meaningful safeguards against abuse in the operation of the Rule, not merely redress against the outcome. For example, there is no transparency in how the FCU operates, no requirement that it provide any reasoning for its decisions, and no requirement that a user is heard before their content is taken down. On the question of bias, it is submitted that there is a difference between an Estate Officer and individuals tasked with adjudicating the legality of speech concerning the Government. The Supreme Court has for decades closely supervised the composition of statutory tribunals precisely because of concerns over independence. Indeed, the Supreme Court has even struck down a constitutional amendment that it found to offer the executive too much influence over judicial appointments. These decisions highlight that the ex-ante composition and independence of decision-makers do matter, particularly where the government itself has a stake in the decision made by the tribunal or bench. As Justice Patel observes,

“There is no safeguard against bias. There are no guidelines, no procedure for hearing, no opportunity to counter the case that some information is fake, false or misleading. […] Even more disturbingly, the Rule clearly makes the Central Government a judge in its own cause. (¶189 Patel J.)”    

Given that Rule 3(1)(b)(v) concerns the falsehood of information concerning the Government, the composition of the FCU undoubtedly matters. The FCU allows the Government to exclusively decide whether information concerning the Government in the public realm is correct or wrong.


These two blog posts have attempted to capture the analytical differences between the Judges on the constitutionality of Rule 3(1)(b)(v). But ultimately this case was also about citizen-State relationship and the role of free speech in mediating this relationship. This case concerned the Government claiming a unilateral right to determine what information in the public sphere about it was true. The Government inter alia argued that false speech about the government was not protected, that the Government acted as parens patriae for citizens in the informational ecosystem, and the Government was fulfilling a sovereign duty to ensure that citizens get the ‘correct’ information about the Government. These are troubling arguments that are antithetical to our understanding of democracy which is premised on citizens freely choosing amongst competing visions of the public good. Justice Patel’s opinion emphatically rejected these submissions:

“I reject without hesitation the attempt to curtail Article 19(1)(a) buried in the submission that the fundamental right is to ensure that every citizen receives only ‘true’ and ‘accurate’ information – as determined by the government. It is not the business of the government to keep citizens from falling into error. It is the other way around. For it is very much the business and should be the duty of every citizen to prevent the government from falling into error. Governments do not select citizens. Citizens elect governments. (¶158 Patel J.)” 

The baton is now passed to the judge(s) tasked with resolving the split verdict.

*Disclaimer: One of this Blog’s editors (Gautam Bhatia) was a counsel in this case. He was not involved with the publication of this post.

This article was first published in the Indian Constitutional Law and Philosophy blog: