This post is authored by Dhruv Bhatnagar
Through an order dated July 19, 2022 (“Order”), Justice G.R. Swaminathan of the Madras High Court initiated proceedings for criminal contempt against YouTuber ‘Savukku’ Shankar. The genesis of this case is a tweet in which Shankar questioned who Justice Swaminathan met before delivering a verdict quashing criminal proceedings against another content creator. Shankar’s tweet on Justice Swaminathan has been described in the Order as ‘an innuendo intended to undermine the judge’s integrity’.
In the Order, Justice Swaminathan has observed that Chief Compliance Officers (“CCOs”) of social media companies (“SMCs”) are obligated to ensure that “content scandalising judges and judiciary” is not posted on their platforms “and if posted, [is] taken down”. To contain the proliferation of ‘anti-judiciary content’ on social media, Facebook, Twitter, and YouTube have been added as parties to this case. Their CCOs have been directed to document details of complaints received against Shankar and explain whether they have considered taking proactive steps to uphold the dignity of the judiciary.
Given that users access online speech through SMCs, compelling SMCs to exercise censorial power on behalf of State authorities is not a novel development. However, suo moto action to regulate ‘anti-judiciary content’ in India may create more problems than it would solve. After briefly discussing inconsistencies in India’s criminal contempt jurisprudence, this piece highlights the legal issues with standing judicial orders directing SMCs to proactively monitor for ‘anti-judiciary content’ on their platforms. It also catalogues the practical difficulties such orders would pose for SMCs and argues against the imposition of onerous proactive moderation obligations upon them to prevent the curtailment of users’ freedom of speech.
Criminal contempt in India: Contours and Splintered Jurisprudence
The Contempt of Courts Act, 1971 (“1971 Act”) codifies contempt both as a civil and criminal offence in India. Civil contempt refers to wilful disobedience of judicial pronouncements, whereas criminal contempt is defined as act(s) that either scandalise or lower the authority of the judiciary, interfere with the due course of judicial proceedings, or obstruct the administration of justice. Both types of contempt are punishable with a fine of up to Rs. 2,000/-, imprisonment of up to six months, or both. The Supreme Court and High Courts, as courts of record, are both constitutionally (under Articles 129 and 215) and statutorily (under Section 15 of the 1971 Act) empowered to punish individuals for contempt of their own rulings.
Given that “scandalis[ing]” or “tend[ing] to scandalise” a court is a broad concept, judicial interpretation and principles constitute a crucial source for understanding the remit of this offence. However, there is little consistency on this front owing to a divergence in judicial decisions over the years, with some courts construing the offence in narrow terms and others broadly.
- Courts should exercise a “wise economy of use” of their contempt power and should not be prompted by “easy irritability” (¶27).
- Courts should strike a balance between the constitutional values of free criticism and the need for a fearless judicial process while deciding contempt cases. The benefit of doubt must always be given since even fierce or exaggerated criticism is not a crime (¶28).
- Contempt is meant to prevent obstruction of justice, not offer protection to libelled judges (¶29).
- Judges should not be hypersensitive to criticism. Instead, they should endeavour to deflate even vulgar denunciation through “condescending indifference…” (¶32).
Later, in P.N. Duda (analysed here), the Supreme Court restricted the scope of criminal contempt only to actions having a proximate connection to the obstruction of justice. The Court found that a minister’s speech assailing its judges for being prejudiced against the poor, though opinionated, was not contemptuous since it did not impair the administration of justice.
However, subsequent judgments have not always adopted this tolerant stance. For instance, in D.C. Saxena (analysed here), the Supreme Court found that the essence of this offence was lowering the dignity of judges, and even mere imputations of partiality were contemptuous. Later, in Arundhati Roy (analysed here), the Supreme Court held that opinions capable of diminishing public confidence in the judiciary also attract contempt. Here, the Court noted that the respondent had caused public injury by creating a negative impression in the minds of the people about judicial integrity. This line of reasoning deviates from Justice Krishna Iyer’s guidelines in Mulgaokar, which had advised against using contempt merely to defend the maligned reputation of judges. Not only does this rationale allow for easier invocation of the offence of contempt, but it is also premised on a paternalistic assumption that India’s impressionable citizenry may be swayed by malicious and irrelevant vilification of the judiciary.
Given the above disparity in judicial opinions, Shankar’s guilt ultimately depends on the standards applied to determine the legality of his tweet. As per the Mulgaokar principles, Shankar’s tweet may not be contemptuous since it does not present an imminent danger of interference with the administration of justice. However, if assessed according to the Saxena or Roy standard, the tweet could be considered contemptuous simply because it imputes ulterior motives to Justice Swaminathan’s decision-making.
It is submitted that the Mulgaokar principles more closely align with the constitutional requirement that restrictions on speech be ‘reasonable’ as the principles advocate only restricting speech that constitutes a proximate threat to a permissible state aim (contempt of court) set out in Article 19(2). For this reason, as general practise, it may be advisable for judges to consistently apply and endorse these principles while deciding criminal contempt cases.
Difficulties in proactive regulation of ‘anti-judiciary content’
Justice Swaminathan’s observation in the Order that SMCs have a ‘duty to ensure content scandalising judges is not posted, and if posted is taken down’ suggests that he expects such content to be proactively identified and removed by SMCs from their platforms. However, practically, standing judicial orders imposing such broad obligations upon SMCs would not only exceed their obligations under extant Indian law but may also lead to legal speech being taken down. These concerns are elaborated below:
Incompatibility with legal obligations:
Although the Information Technology Act, 2000 does not specifically require SMCs to proactively monitor content, an obligation of this nature has been introduced through delegated legislation in Rule 4(4) of the 2021 IT Rules. This rule requires SMCs qualifying as ‘significant social media intermediaries’ (“SSMIs”) (explained here) to, inter-alia, “endeavour to deploy” technological measures to proactively identify content depicting rape, child sexual abuse or identical content previously disabled pursuant to governmental or judicial orders. However, ‘anti-judiciary content’ is not a content category which SSMIs need to endeavour to proactively identify. Thus, any judicial directions imposing this mandate upon them would exceed the scope of their legal obligations.
Further, in Shreya Singhal (analysed here), the Supreme Court expressly required a court order determining the illegality of content to be passed before SMCs were required to remove the content. However, if proactive monitoring obligations are imposed, SMCs would have to identify and remove content on their own, without a judicial determination of legality. Such obligations would also undermine the Court’s ruling in Visakha Industries (analysed here), which advised against proactive monitoring to prevent intermediaries from becoming “super censors” and “denud[ing] the internet of it[s] unique feature [as] a democratic medium for all to publish, access and read any and all kinds of information” (¶53).
Unrealistic expectations and undesirable content moderation outcomes:
Judicial orders directing SMCs to proactively disable ‘anti-judiciary content’ essentially require them to objectively and consistently enforce standards on criminal contempt on their platforms. This may be problematic considering that the doctrine of contempt emerging from constitutional courts, where judges possess a significantly higher degree of specialised knowledge on what constitutes contempt of court, is itself ambiguous at best. Put simply, when even courts have regularly disagreed on the contours of contemptuous speech, it may be problematic to expect SMCs to take more coherent decisions.
A major risk with delegating the burden of complex decision-making about free speech to private intermediaries is excessive content removal. Across jurisdictions, platform providers have erred on the side of caution and over-removed content when faced with potential legal risks. This is evidenced through empirical studies on the notice-takedown regime for copyright infringing content in the US and due diligence obligations for intermediaries in India.
Given their documented propensity for over-compliance, directions by Indian courts requiring SMCs to proactively takedown ‘anti-judiciary content’, may incentivise excessive removal of even permissible critique of judicial actions by SMCs. This would ultimately restrict social media users’ right to free expression.
Considering the issues outlined above, it may be advisable for the Madras High Court to refrain from imposing proactive monitoring obligations upon SMCs. Consistent with the Mulgaokar principles, judges should issue blocking directions for online contemptuous speech, in exercise of their criminal contempt jurisdiction, only against content which poses a credible threat to the obstruction of justice and not against content which they perceive to lower their reputation. Such directions should also identify specific pieces of content and not impose broad obligations on SMCs that may ultimately restrict free expression.