The Impact of Digital Media on the 2019 Indian General Election

Executive Summary

The 2019 Indian general election saw the widespread and innovative use of digital media and technology. Most political parties employed digital media extensively for campaigning and voter mobilisation. This extensive use of digital media was due to nearly half of India’s 900 million eligible voters having access to the Internet and social media. The country has 300 million Facebook users and over 200 million on WhatsApp, which is more than any other country. Indeed, the 2019 election was dubbed by many as the ‘WhatsApp’ election. In addition, millions in India use other media platforms such as the regional language platform ShareChat and the globally popular TikTok.

The Institute of South Asian Studies (ISAS) at the National University of Singapore and the Konrad Adenauer Stiftung (KAS) organised a joint roundtable titled ‘Digital Media, Politics and Elections in India’ on 28 June 2019. The event brought together researchers, policymakers and representatives of digital media companies.

This report largely draws upon the discussions at the roundtable. It focuses on two key areas. First, the new campaign strategies used in 2019, such as different types of political rhetoric, the use of vernacular language mediums and the seamless transition between the online and offline space. Second, the imperatives of regulation and how they relate to artificial intelligence, ethics and state institutions.

There have been remarkable changes between the Indian general election of 2014 and 2019. The digital divide within India has increasingly narrowed and this has made it even more important to study the effect of digital media on politics. In 2014, 1 in 5 Indians had access to digital media and that has now increased to 1 in 3. The quality of access had also dramatically improved with low-priced smartphones and cheap data plans.

The 2019 election also saw a spike in online rumours and fake news. There were efforts at regulation by the government, particularly the Election Commission, and tech companies. These, however, turned out to be of limited efficacy.

Introduction

The use of digital media in the 2019 election campaign was vastly greater than in 2014. The incumbent Bharatiya Janata Party (BJP) no longer had the first mover advantage and it was matched by an increase in digital campaigning across all parties. In the run up to the 2019 election, most parties had a social media wing. The 2019 election also saw individual politicians forming their own social media teams. However, the BJP was still ahead of the curve with reportedly three WhatsApp groups for each of India’s over 90,000 polling booths and 1.2 million social media volunteers. The NaMo app, which tracked every electoral activity of Prime Minister Narendra Modi, was extremely popular too.

The increased use of digital media by political parties was matched by a corresponding increase in the number of internet users in India. According to one report, the number of internet users in India had crossed 550 million in end-2018[1]. The growth has come on the back of a steady growth of cellphone subscribers, which had crossed the billion mark in 2018. The cost of mobile data was another factor too. According to the BBC, 1 gigabyte (GB) of mobile data costs $0.26 in India, compared with $12.37 in the US, $6.66 in the UK and a global average of $8.53.[2] The quality of access has also improved, with the Reliance Jio phone providing 8 GB of data per month at throwaway prices, giving individuals ample data to browse the internet for political content. By December 2018, two years since its launch, Jio had become the third largest telecom operator in India with 280 million subscribers. Each Jio mobile phone also came pre-installed with the NaMo app.  Overall, when compared with 2018, there had been a 21 percent increase in internet users and a 24 percent increase in the number of active social media users (Figure 1). When compared globally, India ranked second on year-on-year change in social media users, with a 24 percent growth between 2018 and 2019 (Figure 2). This figure has almost doubled since the 2014 WeAreSocial report[3].  

One of the questions that needs to be asked is whether the impact of social media on politics has been exaggerated or not. For instance, in South India, which has the highest percentage of WhatsApp, Facebook, Twitter and YouTube users, there was a negative correlation between the actual election results and the expected results based on social media use. In the Hindi belt, the BJP also won in states where they had a lower social media reach. Survey data from the Lokniti report on ‘Social Media & Political Behaviour’ also suggests that the BJP would have won the 2019 election even without the help of social media. The report, however, also says that social media played a vital role in spreading awareness about proposed schemes such as the Congress party’s Nyuntam Aay Yojana (NYAY) and the Indian government’s actions such as the February 2019 airstrike on a terrorist camp in Balakot in Pakistan[4]. Those who were active on social media had more awareness about these than non-users. Therefore, while the role of social media in the eventual election outcome can be debated, its impact on campaigning cannot be denied. What the Lokniti report underestimated was the ‘second order’ effect of digital media, which contributed to the framing of political personalities and the way the offline election campaign was conducted.

New Campaign Strategies

In 2014, the use of social media helped Narendra Modi grow his popularity to eventually be termed, by The Financial Times, as ‘India’s first social media prime minister’. The BJP’s 2014 election campaign used social media to good effect. However, during the 2019 election the BJP no longer had the first mover advantage. Most other parties increased their digital campaign efforts and expenditures. For instance, the Congress employed a research, graphics and professional IT team to counter BJP’s social media campaign. Indeed, the Lokniti report points out that the BJP’s social media advantage over the Congress declined in 2019.

However, the BJP still had the most structured and creative social media campaign compared to other political parties. With multiple layers of communication, a central vision and structure of information flow was implemented. All messages from Delhi had to be forwarded and ministers were expected to retweet Modi’s tweets. The party relied heavily on WhatsApp using a multi-step distribution model of official networks, state units, middle layer intermediaries (bhakts), well-wishers and local influencers. Individual politicians and ministries also built their own social media teams, with each one having a social media coordinator to capture photos from rallies or events and share them online.

The BJP also invested heavily on political outreach in digital platforms. A total of USD 11 billion (SG$15.2 billion) was spent on political advertising in 2019, and 60 percent of the ads were by the BJP. The BJP’s heavy use of social media meant that the party reduced its budget for advertising on traditional media.[5]

Regional Language Social Media Applications

Language has always played a crucial role in election campaigning. At the heart of the 2002 Gujarat Assembly election and the BJP’s eventual win, was Modi’s very clear understanding of how the media functioned. He switched between English language media and regional language media to establish an ‘us versus them’ narrative. The BJP’s 2002 campaign was largely centred on this narrative and contributed to its eventual win.

During the 2019 election, most political parties utilised regional language platforms to gain more traction amongst the electorate. In contrast to earlier elections, regional languages were widely used to connect to voters. A University of Michigan study attributed Rahul Gandhi’s growing traction on Twitter to his use of Hindi.[6] In Tamil Nadu, for instance, the Dravida Munnetra Kazhagam (DMK) launched its own mobile application, which allowed the party to create content in Tamil and share it to workers internally. These workers would then disseminate it to others outside the party.

The launch of Jio minimised the digital divide significantly. As a result, there was a boom in the number of people coming online for the first time, many of whom were looking for a social media network that functioned within their language specifications. This gave rise to new regional content platforms such as ShareChat and Helo that catered to regional languages. Indian language platforms, such as ShareChat and Helo, operating in as many as 15 different languages (Hindi, Tamil and Telugu chief among them), were major sites of election propaganda. Both Helo and ShareChat are now targeting the 100 to 150 million mobile internet users in rural India and Tier 2 and 3 cities that are populated by Indian language speakers. ShareChat had its largest user base in Tamil Nadu and this translated to political parties using the app at a broader level. The DMK created accounts on ShareChat for each district and constituency in the state, which amounted to a total of 300 accounts.  

There has thus been a growing influence of regional language social media applications that are more effective at targeting the local population. The changes in the Top 10 Free Apps in Playstore from 2017 to 2018 reflected this trend. Commonly used apps such as Facebook and Messenger had been replaced in 2018 with more vernacular language apps such as ShareChat and Helo (Figure 3).

Political Rhetoric

The 2019 election saw the use of insults on a large scale, leading many to comment on how low the quality of public discourse was. However, politics has always used name-calling, insults, sarcasm and other elements of political communication that are sensational. The difference in 2019 was the amplification effect through social media. With mobile phones being common and more people having access to the internet, this low quality of public discourse has become routinised, perhaps even the new normal.

This is problematic because the younger voters assume that this is the standard modus operandi for political rhetoric. Name calling or insults have become acceptable because of the ‘disinhibition’ effect[7]. Many assume a vast distance between themselves and their audience, not realising the implications of what they say online. Therefore, with name calling and personal attacks becoming so common, political communication and public discourse seems to have accepted this as the norm.

At present, trust in social media is low and users no longer trust the news that they see on social media. According to a Reuters Institute report, the proportion of survey respondents in India that trust the news in social media at 34 percent is lower than that of those who trust news overall at 36 percent (Figure 4). Similar levels of distrust were reported in the Lokniti report, which found that one out of four respondents claimed to have no trust at all in the news shared on social media platforms.

Political parties have also leaned into this new mode of political rhetoric. The BJP used humour, wit and sarcasm in their digital campaigns and constantly attacked their opposition. This caused a further deepening of the affective infrastructure of the BJP and day-to-day vocabulary, making it very banal. It also helped to make conversations about Hindu nationalism entrenched in the political landscape and discourse. The conversations about the ‘nation’, or ‘nation’ talk, stretched the boundaries of what could be said in public and created an everyday familiarity with nationalist vocabulary.  

The BJP’s agility and large volunteer base on the ground meant that their IT cell had a good gauge of public sentiment and could convert insults into an advantage for the party. This was evident in how the BJP turned the Congress’ ‘Chowkidar chor hai (the watchcman is corrupt)’ slogan, targeting Modi, into their favour by coining a new slogan and hashtag, ‘Main bhi chowkidar (I am a watchman too)’.The BJP was adept at micro-targeting voters and understanding voter sentiment on the ground level, so as to incorporate these insights into their social media strategy. This synergy between the online and offline realms was a significant advantage that the BJP had vis-à-vis other parties.

Online versus Offline

Despite the growing significance of digital media, conversations about digital media presence or digital avatars cannot be entirely detached from offline modes of campaigning.

The seamless transition between the online and offline realms meant that parties needed to understand the importance of pairing digital campaigning with strong action on the ground. Apps such as ShareChat understood this and utilised their micro-level account sharing for individual districts to reach a wider audience than Facebook or Twitter. The close link to the ground level allowed ShareChat to be able to predict the outcome of elections, using likes and comments, with remarkable accuracy. Parties also used digital media to conduct needs-based assessment to analyse citizen needs and grievances. The Aam Aadmi Party (AAP) conducted online audits in order to build their manifesto.

Ignoring these factors limited the efficacy of campaign strategies. This was made evident in the Congress’ NYAY programme. NYAY was a minimum income guarantee scheme targeted at lifting poor families out of poverty. However, it was advertised in very high-end malls with multiplexes, where the intended beneficiaries were not likely not visit. Whilst the digital advertisement was factually accurate, its placement was far removed from the real context.

This relationship worked in reverse as well with offline inefficiencies translating into online limitations. The BJP had a unified stance on their narrative and use of Modi as the face of their party. However, the Congress did not have this unity and their organisational inefficiencies translated onto social media. The Congress was more intent on countering insults to party president Rahul Gandhi and the ‘pappu (inexperienced or naïve person)’ narrative targeted at Gandhi, rather than scripting a narrative of a Congress revival.

Regulatory Challenges

Fake News and Artificial Intelligence

The low levels of trust in social media means that audiences are now realising that the content that parties push onto social media platforms are largely propaganda. This heightened awareness of propaganda and fake news has seen more users reporting content to fact checkers. These users are active and accurate in reporting untrue information to sites such as NewsChecker. User reporting has been more common in Telugu and Hindi, and more limited in Malayalam and Punjabi. The Computational Propaganda Project at the University of Oxford found that between mid-February 2019 and the eve of the first round of voting, in mid-April, “the proportion of polarizing political news and information in circulation over social media in India was worse than all of the other country case studies” it had analysed, except the US presidential election in 2016.[8]

This was especially true during the 2019 terrorist attack in Pulwama in Jammu and Kashmir and the subsequent Balakot airstrike, which occurred months before the general elections, and resulted in a spike in fake news across various platforms. This meant that issues that were originally not associated with the election campaign, such as national security, were now being used to garner votes through fake news. Similarly, fake news had also been spreading in Kashmir, proving to be an issue of local safety, security and stability. Fake news created the feeling that the country was under attack and that citizens needed leadership which would protect them and provide safety. In some cases, fact checkers stepped in to spread awareness on facts and fake news. For instance, after the Balakot airstrike there were fake videos circulating showing the impact of the strike on a terrorist camp in Pakistan. Fact checkers were quick to debunk these claims by pointing out the video was a simulation and not real footage. However, very often the damage had been done and the misinformation had spread widely before it was debunked.

The 2019 election saw the use of artificial intelligence (AI) tools to combat the problem of fake news. The fact checking agency, BOOM, had access to a dashboard that flagged information as false and dug deeper into posts that could potentially be dangerous. It then put a strike on the post and Facebook suppressed the post’s distribution. This resulted in mainstream media facing an immediate decline in traffic. BOOM is presently working with YouTube and Google to improve the legitimacy of news articles. For instance, every search result on a topic is paired with an article written by BOOM on the reliability of that article. This would immediately signal to readers whether the article that they are reading, is accurate. Similarly, WhatsApp put in place a ‘checkpoint research tipline’ that worked together with other companies for users to submit potential misinformation. The tipline then created a database of misinformation during elections and flagged 70-75,000 messages in one month.

However, use of AI faced several challenges. First, political parties created content much faster than AI tools could keep up with. Second, AI tools did not prove very useful for fake news that used visual content. For instance, even if the photos and videos were real, the context might be vastly different, and it required a deeper digging by a human researcher to find out the real context. For instance, a video of actor Akshay Kumar, saying that terrorism was an issue, was from years ago or a statement by Shah Rukh Khan claiming that he would leave India if Pakistani actors were boycotted by Bollywood was never even made. Third, given how big witty and sarcastic insults were in the 2019 election, AI tools failed to catch such remarks and categorise them as insults. This meant that the literary techniques of sarcasm and wit had to be understood by strong qualitative human methods. Fourth, AI was limited in its understanding of regional languages. Again, it must be paired with human understanding of what was deemed as important keywords, before they were fed into the machine. WhatsApp’s tipline also translated reports from English or Hindi to other regional languages, which could deviate from the intended meaning. Fifth, AI also fared poorly in terms of prediction. This was largely because a lot of prediction models built trendlines based on prior assumptions of what voter behaviour was like in previous elections. However, this was not objective and left out new factors that should have been taken into consideration such as changes in voter sentiment.

Role of the Election Commission

India’s powerful Election Commission (EC) serves as a key institution during the election period. In the 2014 general election, the pre-certification of social media content was mandated. Therefore, advertisements had to be certified within the boundaries of permissibility for an electoral speech. They also had to be certified as not appealing to caste or religious identity and as not promoting hate speech or bribery.   

The EC’s Model Code of Conduct (MCC), which is in place during elections, is a mechanism for norm definition and for the commission to promote good conduct. It has served as the primary bulwark for electoral integrity. However, the code lacks in enforceability. Out of the 510 code violations in the 2019 general election, 93 were shortlisted, 75 analysed and some inferences were made. Yet, there was no disqualification or direct legal penalty for violating the code. Thus, the code was substantially deficient to deal with electoral violations.

For the 2019 election, the EC put in place a Voluntary Code of Ethics, agreed upon by major social media platforms, for social media regulation. According to the code, the platforms would endeavour to, “where appropriate and keeping in mind the principle of freedom of expression,” deploy appropriate policies and processes to facilitate access to information on electoral matters. The EC also added social media experts to its district and state-level election Media Certification and Monitoring Committees and launched an app called cVigil through which MCC violations, both online and offline, could be reported. However, its record of responding to complaints was tardy.

A possible future route would be to look beyond the Model Code of Conduct and the Voluntary Code. For instance, in preparation for the 2019 election, the election commission worked with WhatsApp since July 2018. Together, they educated 150 agencies on law enforcement. Such collaborations with tech companies could prove useful in regulating social media during elections.

Question of Ethics

The ethical dimension also needs to be looked at because it has implications on several other aspects of society. There is a large set of ethical issues that arise solely from the way social media was used during this election. There were questions on how political parties and supporters were using and collecting personal data, how advertising and marketing were funded and issues on how algorithms determined news feeds and the information flows.

In looking at ethics in this largely networked age, the significance of digital literacy needs to be emphasised. Given that AI tools can only provide piecemeal solutions to countering disinformation, the population needs to be made more aware. This could be done through the inclusion of digital literacy education from a primary school level. Overall, a coalition of political actors, fact checkers, civil society groups and think tanks, need to collaborate and rethink the way people understand facts.

Conclusion

Since the extensive use of digital media played an important role in the 2019 Indian general election, the roundtable discussed new campaign strategies, their impact on public discourses and questions surrounding ethics and regulation.

The sheer numbers in India are impressive – over 500 million internet users, 300 million users of Facebook and 200 million users of WhatApp. This, however, also means that over half the Indian population still does not use digital media. Television and newspapers remain the main source of political news for most voters. Since it is expected that with millions of new users will have internet access by the next elections in 2024, the importance of social media will continue to increase significantly.

The use of digital media in the 2019 Indian general election campaign was remarkable, both in terms of distribution and content. The highly structured dissemination of messages across multiple layers of communication was notable, as was the strategic use of a diverse range of platforms by the BJP. The party used a top-down system to disseminate messages, which was centrally controlled, via social media and messaging services. This also had an agenda setting effect. Topics that were discussed on social media were taken up by traditional media.

In this context of the increasing use of digital media by political parties, leaders and candidates, digital and network literacy needs to be emphasised. To counter the spread of fake news, it is important to further sensitise citizens and celebrities in order to prevent the forwarding of such messages in future campaigns. Political actors, tech companies, fact checkers and civil society groups need to collaborate in this area. The ‘checkpoint research helpline’ that was initiated by WhatsApp or the collaboration of Google and YouTube with fact checkers are important steps in this regard.

An increase in nationalistic or exclusionary language, often through humour and jokes, could be observed. This ‘nation talk’ stretched the boundaries of what could be said and might have contributed to everyday familiarity with nationalist vocabulary. Furthermore, with name calling and insults becoming quite common, public discourse seems to have accepted this as the ‘new normal’.

The impact of digital media campaigns on the actual election results in India cannot be quantified. Electoral outcomes are not the only barometer though: More research is needed on the ‘second order effects’ of social media on offline campaigns, traditional media, public discourse and framing of political personalities. In future, political parties will rely even more on the use of social media for campaigning and voter mobilisation. In order to combat the negative effects of fake news or low quality of public discourse and to strengthen the potential for informing voters, a new multi stakeholder dialogue is needed. Cooperation among political parties, technology companies and civil society, as well as the involvement of the Electoral Commission, is essential.

[1] The Economic Times. (2019). Internet Users in India to Reach 627 million in 2019: Report. https://economictimes.indiatimes.com/tech/internet/internet-users-in-india-to-reach-627-million-in-2019-report/articleshow/68288868.cms. Accessed on 16 September 2019.

[2] Roy, P. (2019). Why India has world's cheapest mobile data. BBC News. https://www.bbc.com/news/world-asia-india-47537201, Accessed on 16 September 2019.

[3] Simon Kemp, ‘Digital 2014: India’, 2014. https://datareportal.com/reports/digital-2014-india

[4] LOKNITI, Centre for the Study of Developing Studies, 2019. Social Media & Political Behaviour. https://www.csds.in/uploads/custom_files/Report-SMPB.pdf.Accessed on 9 September 2019.

[5] Anumeha Chaturvedi, ‘BJP top spender on political ads on digital platforms’, The Economic Times, 16 May 2019. https://economictimes.indiatimes.com/news/elections/lok-sabha/india/bjp-top-spender-on-political-ads-on-digital-platforms/articleshow/69351792.cms?from=mdr .

[6] Mandira Banerjee, ‘Tweeting in Indian language widens polictians reach’, University of Michigan News, 19 September 2019. https://news.umich.edu/tweeting-in-indian-language-widens-politicians-reach/

[7] The disinhibition effect refers to the loosening of social norms and etiquette that are normally present in a face-to-face setting, in interactions online.

[8] Allsop, J. (2019). Results expected in India’s ‘WhatsApp election’. [online] Columbia Journalism Review. Available at:  https://www.cjr.org/the_media_today/india_election_facebook_fake_news.php  Accessed 16 September 2019.