India Needs a Fresh Strategy to Tackle Online Extreme Speech

Allegations and counter-allegations around abusive speech and disinformation on social media networks have made “trolling” and “fake news” significant actors in public discourse, as India braces up for the next round of national elections.

From serious allegations around paid trolls to casual-jocular naming of an irritating user as a troll, brazen language, “fact-filled” untruths and belligerent tone of exchange have become an everyday reality of online political debates. With several prominent online users coming out in the open to complain about abusive speech, online harassment is no longer a private injury that is dealt with in a hush-hush manner—with the nervous, invisible clicks of delete and block buttons. With several new civil society and business initiatives launched to do fact-checking, the quality of information exchange has at the same time become a major concern. Online abuse and disinformation have struck the public mainstream, bringing the spotlight on the “dark side” of internet exchange.

While global concerns over fake news and disinformation have found resonance in the world’s second largest online sphere (with 450 million internet users), social media exchange in India should also be understood in relation to longer cultures of political exchange and structures of privilege that define who gets to spread disinformation or abuse fellow users, and with what consequences. Far from dismissing rugged exchanges on online media as political mud-slinging common to India’s fissured democratic landscape, it would serve well to approach abuse and disinformation as an important culture of mediatised politics in the digital age, which not only reflects extant political differences but significantly shapes what it means to participate in public life for a net fed generation.

“Extreme speech” is a conceptual tool that could help to grasp the phenomenon of online abusive exchange and disinformation. Extreme speech refers to speech acts that stretch the boundaries of acceptable speech along the twin axes of civility/incivility and truth/falsity. As a concept, “extreme speech” resists the temptation to label all manner of vitriol as “hate speech”. While regulatory excess that could arise out of such overreach is obvious, what is equally important, from a theoretical point of view, is the ambiguity of public speech forms. Mazzarella and Kaur (2009) capture this as “cultural regulation.”

Rather than a field of polar opposites, cultural regulation defines a “spectrum of public cultural interventions” between censorship and publicity that feed on each other to “generate value (commercial or symbolic) out of a delicate balancing of incitement and containment.” As regards online speech, this explains the contradictory climate of indulging in online vitriol and its public disavowal, which is itself reflective of the broader tension between containment as a value in the regulatory domain and incitement as an attribute of practical politics. In other words, no political party or vested group will be untouched by the culture of online vitriol and disinformation in the days to come.

From an analytical perspective, extreme speech allows us to pry open the field to map different forms, formats and actors involved in online vitriol and disinformation. A departure from the overarching category of “hate speech” allows for closer attention to actual user practices and different new media affordances that enable such action. For instance, I have argued that online abuse in India could be best captured as “gaali” which signals the blurred boundaries between comedy, insult, shame and abuse emerging on online media with divergent consequences (Udupa 2017). Such grounded analysis approaches online abuse as “extreme speech” representing a spectrum of online practices rather than the culturally flat antonymous conception of uncivil speech versus acceptable speech (Pohjonen and Udupa 2017).

The key differentiator is that there is a performative moment to transgress the norm, and incentives found in online media and political ecosystem to say the unsayable, make jibes where none was permitted, or creatively mash up messages to offer a distorted view. In all these moments, the “mainstream” or the “norm” are themselves moved and readjusted. While such acts can have a subversive potential to “talk back to authorities,” they can also be a tool for dominance when used from positions of privilege.

How do we use these perspectives to map online extreme speech in India, which is gaining momentum with upcoming national elections as a critical event? The brief analysis that follows is based on ongoing ethnographic fieldwork among politically active online users in Mumbai, Delhi and Bengaluru that started in 2013, and social media content and network analysis of online exchange based on purposive sampling. Methodologically, it is rooted in “internet related ethnography” combining onground fieldwork with discourse and network analysis of social media exchange (Pink et al 2016).

Form, Formats and Types

In our ongoing research on digital politics, we have found that extreme speech is proliferating across social media platforms. Far from an ahistorical position that would consider these forms as a new media phenomenon, such expressions should be understood in relation to cultures of political sloganeering, subversive speech, and efforts to “semiotically dominate the opposition” via oratory (Bate 2009) and campaignstyle manoeuvring in postcolonial India. This history is beyond the scope of this short essay, but suffice it to say that online media have brought these styles to the fore of everyday political engagement, infusing them with globally resonant online user cultures.  Online extreme speech reflects and transforms face-to-face interactions set within polymedia environments where television channels and a section of print media are increasingly embracing accessible and provocative language.

Although extreme speech expressions are seen across social networking and messaging services, it is possible to differentiate between platform-specific and platform-agnostic formats of extreme speech.

Memes: Internet memes, for instance, cut through different social media platforms, as they are shared on Facebook, Twitter, Instagram and WhatsApp, depending on the popularity of specific platforms within specific communities. In ethnographic interviews, we have found vast variation in the use of these platforms for political discussions. While online actors more comfortable in regional languages preferred Facebook and WhatsApp to Twitter, urban users fluent in English found Twitter to be a “serious forum” for engaged political commentators. In “assigning political moralities” to social media platforms (Miller 2011), user groups differed in their preferences, reflecting linguistic differences and class location. At the same time, political parties have started to nurture all these platforms, allowing different platforms to resonate with different target groups. Internet memes available across these platforms are images filled with wit and sarcasm, where “political, social and playful purposes exist simultaneously” (Miller 2011: 70-73). Its “extreme” nature lies in the wilful and playful disregard for sanctioned forms of political discourse or polite (elite) ways of political messaging.

Labels: The second form of online extreme speech is in the labels assigned to online accounts, pages and profiles. These labels range from soft-touch sarcasm to aggressive derision of the opposition. Facebook groups such as The Illogical Indian @illogicaldesi (64,207 followers)^[1], The Frustrated Indian (1.1 million followers)^[2], India Against Presstitutes (455,107 followers)^[3] and Twitter handles such as Eminent Intellectual, @padhalikha, @UnSubtleDesi and @Sussuswami are some example.

Hashtags: The third variety consists of hashtags, where extremeness is in the label but also more fundamentally in the coinage of the hashtag itself. Hashtags are not just about framing prevailing issues, but they actively bring them into existence. In many cases, they instigate the discourse. Some of the most pertinent examples for extreme speech are available in the analysis of hashtags on Twitter and Facebook. One example is the hashtag “gharwapsi” that started to trend in 2015, tagging on to the Hindutva campaign to draw Muslims and Christians “back to the Hindu community.” Reflecting the campaign on the ground, the hashtag provoked voluminous discussions on Twitter around a seemingly pithy expression. Another example is the heated public debate around Jawaharlal Nehru University campus in 2016, after allegations of anti-national slogans stirred a nationwide debate. Hashtags such as #ShutDownJNU and #CleanUpJNU fired up discussions online, in as much as they fed heated clashes on the campus. Such event-centric hashtags gain momentum with mass media coverage and recede as organised media debates fade out (Figure 1).

Expressions with Syntax: The fourth format is verbal expressions with a more complete syntax, sans images. Here, evocative phrases create the effect of “wounding words” (Butler 1997). An example would be lengthy diatribes on Facebook in the description of the pages, posts, and comments. Some of them directly index their emphasis on parody and sarcasm. Facebook page “The Illogical Indian” @illogicaldesi, for instance, states this under “about”: “SARCASM is inevitable, since beating the shit out of hypocrites, in public, is illegal… Hence we counter ILL-LOGIC of so-called LOGICAL pages with sarcasm”.^[4]

Remix: The fifth format is visual remixing, digitally enhanced videos and “unofficial uploads” of grainy video captured on mobile phones. These visual cultures as “small frame politics” are prominent on YouTube (Dattatreyan 2017). They have upset official law and order narratives by offering bottom-up witnessing (Sengupta 2012) but in other instances they have played a key role in spreading rumours.

Across these online formats (not necessarily an exhaustive list), one could draw a typology of extreme speech combining the axes of style and content (the two axes can be kept separate as well, but here I have combined them to indicate the composite nature of extreme speech). The first type involves ad hominem comments and name-calling that target the attributes of the person, often in derisive hilarity (For example: “Kejriwal wears a 42 size shirt when his size is 38”).

The second type is gender-based shaming, based on tropes of sexual modesty and heteronormative order, and related allegations of moral debauchery expressed as illicit sexual relations and promiscuity. A common example here are allegations of female political commentators “sleeping” with party leaders. There are examples aplenty from online debates, and a more recent one we confronted in our fieldwork was the claim that a social media strategist was given a prominent position in a national political party only because she had an illicit relation with the head of the party.

The third type is anti-minority speech, most prominently anti-Muslim messages. A recent example comes from an excerpt from our fieldwork, when our online interlocutor averred “There are Maliks in Pakistan and Maliks in India. Indian Jats follow Hinduism, a worthy religion and Pakistani Maliks are terrorists, without any sense of ethics.” Such messages commonly label Indian Muslims as traitors, mischief makers, Pakistanis or traitors.

The fourth type are rhetorical appeals to national feelings and patriotism, and vehement accusations of being anti-national. Regional nationalism and anti-caste politics complicates this formation, as was evident in the Jallikattu campaign in Tamil Nadu (Cody 2018). In any case, Indian national identity is a resonant trope, with commonly occurring themes of foreign-funded biased media and pseudo-liberal intellectuals, that mingle with tropes of making India a global power, and Hindutva stock topics of cow protection and Hindu civilisational glory. The fifth type consists of caste slurs, especially anti-Dalit messaging; while the other types are rumour, counter-facts and disinformation (debunked later on by organised media or fact-checking organisations); physical threats (death, rape and abduction); and wordplay (derogatory acronyms, sobriquets, monikers for opponents: “commies”, “urban naxals”, “libtards”, “presstitutes”, “Pappu”, “Feku”, “bhakt”, “Gappu”, “Sanghi”).

More often, extreme speech is of hybrid types in which anti-Muslim messages combine with gendered shaming or anti-Dalit speech, or accusation of “anti-national” ridicules political opponents or journalists by referring to personal qualities and attributes rather than argument (ad hominem attacks) or with name-calling and labeling.  

An example for the first type is a tweet posted on 1 October 2017 in a tagged response to Hindustan Times report that had tweeted, “Pakistan minister shares stage with LeT founder #HafizSaeed ‘on Pak PM’s instructions’”. In response, the tweeter quipped: “Big deal, its just stage, many Indian journos want to share bed with him…(sic).”

Another example is the accusation that Dalit political leaders are “Muslim sympathisers” because they want “Muslim votes.” An example for the hybrid form of anti-national allegations comes from the tweet pool for the hashtag #JNU in which labeling (as “traitor”) was seen in the top 10 retweeted posts.

An excerpt below suggests its resonance:

RT: #JNU #JadavpurUniversity #HyderabadUniversity all r result of Manipulative Politics;Traitors planning Smthg Big,Kick thm …(321)

RT: KamleshKumari sacrificed her life in 2001 parl attack which hero of #JNU Afzal Guru had plotted #ArrestTheTraitors https://t.… (256)

RT: अब तो भी आतंकवादी सोच रहे होंगे कि हम कश्मीर के लिए फालतू में कुत्तों की तरह मर रहे है जबकि भारतीय गद्दार फ्री में देने…(162)^[5]

In the same hashtag tweet pool, the most retweeted post was the allegation that JNU students had “disgraced” free speech provisions (1039 retweets), followed by the post that read: “Pappu takes U-turn..Now he supports Police Investigation in #JNU So Before Lion @Swamy39  enters into fight..Rat takes U-turn” (813 retweets; note the wordplay, “Pappu”).

One of the effects of extreme speech across these formats and types is the divisive cacophony in political debates, often precipitating into two die-hard positions of imagined liberals versus Hindu nationalists. While in reality these groups overlap and different interest groups intersect, extreme speech pushes these positions to be starkly visible as warring factions, in turn cementing the logics of Hindu nationalism as everyday common sense.

Based on textual content and network analysis of theoretically sampled Twitter hashtags (with a focus on religious nationalism and hashtags derived organically from ethnographic fieldwork as well as through media events), we have found that a set of highly prominent actors drive discussions online. Clusters of discussions surrounding prominent tweeters is a common pattern across the 10 hashtags we have analysed so far (#gharwapsi; #ModiInsultsIndia; #ModiIndiasPride; #JNU; #ShutDownJNU: #CleanUpJNU; #Kathua; #Unnao; #LetsTalkAboutTrolls; #GurmeharKaur). A striking example is the hashtag #kathua (Figure 2) following the brutal rape, abduction and murder of a minor girl who belonged to the nomadic Muslim community of Bakarwals in Jammu and Kashmir (Times of India 2018).

Visualisation based on retweet networks^[6] which revealed two polarised clusters with a few highly prominent users (the two top tweeters by Eigenvector centricity measure that shows most influential nodes in the pool have 62,000 and 134,000 followers respectively as on 17 December 2018). Visualisation based on retweet networks revealed topic polarisation between tweeters defending the Bharatiya Janata Party or those critical of anti-BJP voices on the one hand, and those critical of Hindu nationalist politics or more narrowly, BJP’s specific role in the case.

Aside from revealing the prominence of power users (who come from formal politics, mainstream media or rose to become one), such graphs reveal that real life events are not only reflected but also augmented by social media discussions. Exclusionary extreme speech voicing majoritarian concerns is a common occurrence within these social media exchanges. No doubt, Hindu nationalism is fiercely contested on online media, and the shifting climate of electoral gains reveals greater flux in online discourses. It should be noted that extreme speech as a subversive practice is also common within Dalit and feminist online spheres. However, the very bickering on social media through means of extreme speech has shifted what counts as the center or the “norm”. Analysis of online debates in our sample, for instance, has offered a few counterintuitive results. Hashtag #ModiInsultsIndia revealed that even the RSS (or Modi in this case) was not national enough for some staunch nationalists online:

“Sanghis sided with the British Sanghis killed Gandhi Sanghis abuse Nehru,Ambedkar Never Raise National Flag till 2002 Now #ModiInsultsIndia”

“So now I get why RSS didn't hoist tiranga for 55 years! They were ashamed of being an Indian. Simple. #ModiInsultsIndia”^[7]

The conjecture that opposition parties (Aam Aadmi Party or Indian National Congress supporters) partly drove this discussion is palpable (since one of the prominent tweeters in this hashtag pool was a Congress spokesperson), but what is notable is the pressure to be identified as a nationalist, and the need to interlace it with tropes of Hindu civilisational glory and Hindu religiosities.

Conclusions thus far suggest that banalisation of exclusion through formats of humour, sarcasm and rumour have been key in cementing Hindu majoritarian positions within India’s online discourse. While direct intimidation in online speech is also prevalent, it is largely through argumentative confrontations and rebuttals that key tenets and tropes of Hindutva have settled as a familiar ideological vision for a new generation of supporters. On the other hand, remixing, mash-ups and fact-filled untruths of rumor mongering as a form of extreme speech have provoked disturbing incidents of physical violence on ground (Narrain 2017). This then brings the focus on the kinds of actors who compose extreme speech, and patterns of political support and voluntary work that have emerged in the last decade of social media expansion.

Actors of Extreme Speech

Ethnographic fieldwork suggests a layered and variegated landscape of actors who compose Hindu nationalist online discourse, and extreme speech in particular. They comprise at least five prototypes.

Techie turned ideologue: A diffused group of highly motivated, English educated and tech-savvy social media volunteers whose knowledge of internet media as adept coders or end users, is key to their claims to public discourse. In many cases within this subset, ideological affiliation is an effect of media practice. A common refrain from these actors is that they do not take any money from the political party, and their work is entirely voluntary. One volunteer took pains to stress how they don’t expect any money for their online voluntary labour, and that it is “pure passion, kuch karna he [we should do something], we are here to change the youth.”

Party worker online: Committed party workers who took up online media as a potent vehicle for communication and mobilisation. Among these actors, ideological affiliation precedes online uptake. The top-down organisational social media structure of the party influences their decisions and networks, as information is handed out and allowed to spread through state, regional and local channels. Local units are actively encouraged to build their own content. BJP and Congress have set the goal to have at least one local activist for every booth as a node for message circulation, in preparation for the national elections. For instance, in Karnataka, A N Nataraja Gowda, head of the IT Cell of Karnataka Pradesha Congress Committee, has said that the party aims to have one “digital youth for every booth” who will build local WhatsApp groups on which “centralised number” will be added (Shimladka 2018). Through these channels, the party aims to send messages via text, video and animations. The plan is to have both “broadcast WhatsApp groups” for centralised messaging and “organic groups” for bottom up messaging. According to Gowda, the party currently has 2000 to 2500 “organic WhatsApp groups” in favor of the party. The local BJP unit too aims to build on existing social media party networks and organise more WhatsApp groups as they gear up for the national elections. Amaresh K, Karnataka BJP IT Cell Convener, in a media interview averred that “there can be no elections without WhatsApp” (Shimladka 2018). In a personal interview^[8], Amit Malviya in New Delhi summed up BJP’s social media strategy.  “We have used various tools at our disposal,” he described, “to make communication more precise, more informative and more consumer friendly so that the ordinary individual, the man on the street, can appreciate what you are saying.” Giving an overview of the IT operations of the party, he said the national team in New Delhi had diversified the online communication channels, “…in the sense that we have gone down to [regional] states and we are telling them, look, we have to have multiple layers of communication, we have to communicate at the central level, we have to communicate at the state level and perhaps localise it even further.”

Such strategies were pioneered by BJP and only partially by AAP in the last general elections, but now many more political parties, including INC, are scrambling for online influence through dedicated networks of party workers.

Political intelligence consultant: includes “politically agnostic” paid digital influencers who offer digital campaigning, data analysis and digital influence as a “business solution”. While international agencies such as Oligvy&Mather recruited for the Modi campaign in 2014, and high-profile influencers such as Prashant Kishor are known in the public domain, the sphere of digital influencers is far more variegated in recent times, with many “data solutions” agencies offering influence enhancement services to interested political parties.

During ethnographic interviews in Bengaluru and Mumbai, such actors who had started their own small enterprises complained about the “lack of awareness” about digital influence services within political party circles and acts of disrespect politicians meted out to them. These actors shared their concerns about the financial viability of their enterprise, even as they carried theirPowerPoint slides in their bags and sought appointments with party officials to convince them to invest in “analytics and data solutions” for political campaigning.

Monetised Hindutva: Small media businesses that actively appeal to online Hindutva supporters for commercial gains and boost online traffic to their pages that could later be monetised with advertisements. These actors actively follow and court Hindutva supporting online users and encourage them to join or follow their pages.

Bhakt business: Arguably the most intriguing group of online workers more likely to engage in extreme speech.  These comprise small-scale social media entrepreneurs who are excited about Hindutva and Modi, and do not hesitate to make a small business out of it. In some cases, these decisions have been driven by livelihood options. One such entrepreneur in Laxminagar in Delhi started the conversation by declaring how he was a staunch Modi fan, but as the interview progressed, he said his digital influence work for a local BJP leader started because of sheer compulsion of gaining the first contract for his just launched digital media company. As his business grew in strength, he said, his adoration for Modi gained more confidence. In these instances, idolatry, ideological affiliation and commercial interests enter a win-win relation. Such actors rarely have an adequate sense of the party social media structure. One of them residing in Bhiwani in Haryana remarked that the party’s social media structure was still a puzzle, even though he has been working for a local “IT Cell.” “This is regulated at a very internal level,” he said, “We are not allowed to discuss these things, but broadly, the central cell sends instructions to local offices for recruitment. Through these local channels, people come in contact with BJP and then they are provided with small monetary incentives to the tune of two to three thousand rupees.” He continued that once websites and Facebook pages are ready, the party does not give any more cash rewards. “But they help us to realise some revenue through these [social media] platforms by making websites or portals.” Monetary incentives are not the only reason after all. “As we get to know people at local level,” our interlocutor said, “It helps in getting jobs for my friends and family.”  “As you rise in position and influence, the sources of revenue increase and you get many more opportunities, both offline and online. Both financial and power-based opportunities, you know how it works in India”, he laughed, hinting at local networks of patronage that he was able to activate by directly managing 10 Facebook pages for the party, aside from overseeing several more pages across “North India” covering Uttar Pradesh, Delhi, Rajasthan and Madhya Pradesh. 

Disincentives and Awareness

In a pioneering study of political trolling in Philippines, Jonathan Corpus Ong and Jason Cabanes (2018) conceptualise “networked disinformation as a distributed labour of political deception to a hierarchy of digital workers.” They argue that “chief architects of networked disinformation” who intimidate dissenting voices and craft new prospects for political clients via digital influence are themselves “precarious architects of precarious labor arrangements in the creative industries that make workers vulnerable to slipping into the digital underground.” The Indian case of five prototypes of extreme speech actors reveals a similar variegated scenario of precarious labor, political patronage, and opportunism.  In addition, there is ideological activism that is independent of financial motivation or economic necessity. One can turn into the other –for example, a techie ideologue may become a full-time party worker online or start a digital analytics business for political parties. In recent ethnographic follow-up interviews, we have found some disgruntled techies coming out of the BJP fold. Regardless, the basic structure of difference defining actors remains the same. While a techno savvy urban user is more likely to have a stable job and offers voluntary work, bhakt business is more precarious. This calls for paying attention to vulnerabilities and structures of disprivilege that have impelled a large number of social media entrepreneurs to disregard the political fallout of their extreme speech labor. It also draws attention to privileged middle-class online users who engage in partisan ideological discourse, experiencing the collective pleasure of “defeating” their ideological rivals online.

What is significant is that different political parties are rushing to adopt a similar digital media influence and analysis architecture. Optimistically, this could be seen as the dilution of ideological hegemony, since there is lately more clamor online. However,  the nature of extreme speech discourse and patterns of labor that go behind it, warrant a far more serious intervention than leaving it to the (digital) “market place of ideas” or political wrangling. Govindraj Ethiraj, promoter of Boom Live, a fact-checking portal, has said in a recent interview that “Most …fake news is originating from people aligned with the BJP. People aligned with the Congress have not propagated fake news with the same frequency yet” (Manish 2018). Nupur Sharma, Chief Editor of OpIndia, another fact-checking initiative, contends that fake information is peddled by “all kinds of political and ideological affiliations”.  

Mindful of possible public criticism and international scrutiny, social media technology companies have taken steps to stem negative effects of extreme speech in the current election scenario. WhatsApp’s recent decision to limit the number of forward messages to five chats (Indian Express 2018a) is an effort to prevent mass forwards from instigating rumors and violence, while its “research grants” is a further attempt to gain insight and possible soft capital (Indian Express 2018b). These steps are welcome. However, a more “process oriented” approach to extreme speech, to follow Ong and Cabenes (2018), needs a concerted co-regulatory effort where digital influencers are actively disincentivised from channelising extreme speech.

What is worrying is that fake news busting as a growing civil society and business enterprise is itself following the patterns of political and ideological fissures. OpIndia, for instance, has declared openly that they do not claim to be “ideologically neutral”, and that they will “continue to be right-leaning” (Sharma, 2018). As opposed to heavily funded fact checking initiatives, grounded, community level interventions would work well to fend off ideological heavyweights who are backed with financial power.

In an agenda-setting concept note on online hate speech, the European Commission emphasised the “importance of supporting civil society organisations and enhancing the development of positive narratives and critical thinking” (European Commission Presidency Conference 2017). It observed that while cooperation between governments and civil society organisations is necessary,  “cooperation of IT companies with civil society also plays a fundamental role in counter-narrative campaigns”. According to the note, there is evidence to show that deploying social media advertising tools to ‘target audiences’ improves awareness and engagement, and leads to a substantial increase of NGO’s social media presence”.

A different experiment is seen in Germany. Inspired by the US website “Monetising the Hate”, German startup hatr.org has developed a business model to monetize online abusive comments. The site sells advertisements by placing them beneath hateful comments. Users who receive such messages are encouraged to pass them on to the site. The website in turn publishes them and sells ads to place them beneath the messages. Users, on their part, may feel relieved for having passed the offensive content to a site where the “trash” gets converted into “cash”. Money collected thus is donated to “feminist and gender projects,” according to the owner of the startup. These initiatives resemble offline efforts to turn hate into charity, when for instance, groups protesting neo-Nazi movements raised money for every meter covered by a recent Neo-Nazi march in a Germany city.

What is needed in India is a multi-language initiative that can draw support from existing voluntary and non-governmental organisations and equip them with tech ready interventions. These interventions include counternarratives (using platforms such as WhatsApp groups), reporting extreme speech cases to companies, and publication of reports for awareness-raising. Regional language portals can act as local interlocutors to prevent scenarios of “oxidising” extreme speech through repetitions  and excitement of exchange (Phillips 2018). This is especially important in combating false information. As regards online abuse, teaming up with local parody and humour groups will be helpful.  Advocacy groups can publicise trolls and counter them by rhetorical flashes on social media, edging away the shaming aspect of abuse.

Retrograde, false and belligerent messaging with political motivations, I suggest, can only be met with counter-speech and more speech – with concerted digital advocacy, awareness, and public detoxification of online abuse – rather than a legal heavy hand. A multi-throng strategy to online extreme speech is urgently needed, as India prepares to face another election, with a new arsenal of digital tools at its disposal.