Sanjukta Pandey quit her job as a hair and make-up stylist in March to devote herself to Indian prime minister Narendra Modi’s re-election campaign on social media.
Ms Pandey, a feisty 32-year-old wearing huge hoop earrings, neon pink lipstick and a tattoo of Mr Modi’s name on her left forearm, now spends her waking hours spreading his election message on WhatsApp and other social media apps. “I’m online almost 24/7. I don’t go to sleep; we want Mr Modi to come back,” she says. “You won’t see anyone getting inked with Rahul Gandhi’s name.”
India’s ruling Bharatiya Janata party is using WhatsApp to wage one of the world’s most sophisticated digital political campaigns, carried out by a vast army of volunteers like Ms Pandey, who are devoted to Mr Modi’s brand of Hindu nationalism.
While campaigns used to be conducted on TV and at large rallies, WhatsApp has become the central battleground of India’s election, which began on April 11 and will conclude on May 19.
The Indian contest follows a divisive election in Brazil, where far-right candidate Jair Bolsonaro swept into power in October — helped in part by a wave of toxic rumours and misinformation, much of it spread over WhatsApp.
Now India is becoming the latest test case of the capacity of the messaging app, whose millions of small groups of encrypted users are often beyond the purview of electoral authorities or independent fact-checkers, to potentially shape the election in a large and important democracy.
For every supporter who says the app has helped bring together families and friends with a cheap communication tool, there are as many critics who fear it has become an impossible to monitor conduit for fake news.
“WhatsApp is the echo chamber of all unmitigated lies, fakes and crap in India, it’s a toxic cesspool,” says Palanivel Thiagarajan, an elected official and head of the IT department of DMK, a regional party in the state of Tamil Nadu who is running against the BJP. “If it were up to me I would say just cut it, there are hundreds of substitutes.”
The messaging app, which has 1.5bn users globally, has risen to popularity particularly outside the US in countries where its parent company Facebook is hoping to grow new revenue streams.
Claire Wardle, a research fellow at Harvard University and co-founder of First Draft, a non-profit group addressing misinformation on social media, says WhatsApp took off with the explosion of smartphone users in countries such as Brazil, Nigeria and India, where it has become “a primary source of information”.
“These questions about its role in the spread of misinformation are not just to do with elections,” she says. “It’s about WhatsApp’s role in societies, full stop.”
In recent years, it has been Facebook itself which has attracted most of the criticism around the spread of false news and electoral manipulation. The report by special counsel Robert Mueller outlined the extensive efforts by Russian actors to manipulate the 2016 US presidential election using Facebook. It also came under fire when it emerged Myanmar’s military was using the social network to incite violence against the Muslim Rohingya minority in the country.
But it is WhatsApp, which Facebook bought for $22bn in 2014, which has become the communications platform of choice not just in India and Brazil, but also across swaths of Europe including Spain and the UK. Mark Zuckerberg, the Facebook founder, has said WhatsApp’s intimate form of communication is the future of the Facebook group.
“In the last year or so we have seen a move from Facebook news feed to more private channels, including WhatsApp and Messenger, particularly in places like Brazil,” says Ms Wardle.
Its encryption system means that in contrast with Facebook or Twitter, WhatsApp conversations are impenetrable even to the company itself, say executives. But that has made it more vulnerable to misuse, especially in elections, say critics, who argue it has become a platform for spreading campaign-related misinformation.
This risk first came to a head in Brazil last year, in what became known as the first “WhatsApp election”. With 120m WhatsApp users in a country of over 211m, the platform was flooded ahead of the October vote with false rumours, doctored photographs and audio hoaxes — much of which helped Mr Bolsonaro. Researchers studying 100,000 images circulating in 347 groups found that only 8 per cent were “fully truthful”.
“Misinformation was huge in Brazil. It was an election plagued with fake news that left behind a country split in half by hatred,” says Fabrício Benevenuto at the Federal University of Minas Gerais and a researcher on the impact of the social media network. “The political discussion ended up being reduced to a meme.”
In India, the BJP has been the most active of the main parties in trying to use WhatsApp to win votes. “I’ve been trying to reach every household via at least WhatsApp,” says Punit Agarwal, the BJP’s social media co-ordinator for the Delhi area.
Mr Agarwal says the party has 74,000 volunteers tasked with spreading its message over WhatsApp. “There was a limited audience last time,” he says. “This time we have a vast audience.”
WhatsApp has become the platform of choice for politicians because of its massive reach that goes beyond a party’s loyal voter base, but also because of the lack of gatekeepers. Messages forwarded through the system have no context about where they originate, but benefit from the trust of coming from a contact.
Mr Agarwal denies that the BJP is spreading polarising content, but public WhatsApp data collected by analysts and anecdotal evidence shows that Indians are being flooded with propaganda memes, much of it anti-Muslim and critical of the opposition Congress party. “WhatsApp groups are considered the most dangerous,” says SY Quraishi, India’s former election commissioner. “The disastrous potential of this media is very strong; you’ve seen how rumours floating [around] can cause havoc.”
Due to the extensive political participation on WhatsApp in India, the company said it began to plan its election strategy early. “We know political parties are using WhatsApp to organise, and we decided to do a test run [to monitor it] during the Karnataka election,” says Carl Woog, WhatsApp’s head of communications, referring to regional elections last May.
At the time, WhatsApp discovered that one of the political parties, which it declines to name, had created a large number of groups using the party’s name all at the same time, and was adding several people to them in an obvious effort to spread propaganda in contravention of WhatsApp rules. “We had a pretty good sense of what was going on and we banned those groups,” he says, adding that this was the first time the company had observed this viral group behaviour.
Kiran Garimella, a researcher at the Massachusetts Institute of Technology who is studying misinformation in India, analysed more than 5m WhatsApp messages posted in 5,000 public groups over the past five months, covering roughly 1m people.
“We have observed that it is specifically focused on image-based, subtle misinformation,” says Mr Garimella, giving an example of doctored screenshots from a reputable news channel.
The top shared images accused Congress leader Rahul Gandhi of being a “fake Hindu” and invented a link to Vijay Mallya, India’s fugitive billionaire businessman. A recurring favourite message shows Mr Gandhi praying at a mosque, insinuating that he is a Muslim to deter Hindu voters.
Similar to the Brazilian campaign, academics say WhatsApp is being used by parties to polarise citizens by playing on communal tensions.
“I’m a member of over 100 WhatsApp groups run formally or informally by BJP supporters,” says Soma Basu, a fellow at the Reuters Institute of Journalism and an Indian journalist. “I’m analysing 75,000 messages I’ve received in this time, many of which are very disturbing and violent, such as a video of a youth being beheaded in the name of religion.”
Religious propaganda is not just a WhatsApp problem and is spreading through Facebook, Twitter, Indian social media app ShareChat and other platforms but, “WhatsApp carries the credibility of the sender, it’s more private and personal, so a lot of things that can’t be said on Twitter or Facebook can easily be said on WhatsApp”, she says.
The study of WhatsApp political groups in Brazil revealed an elaborate “pyramid” structure of how misinformation spreads on the platform, cascading from regional and local activists to individual citizens and their friends. In India, the structure of WhatsApp political groups is also layered and complex.
The BJP’s social media department targets undecided voters with tailored messages, customised according to voting history, class and caste, says a former data analyst for the party, Shivam Shankar Singh. The messages almost never came from official BJP channels, but WhatsApp groups organise outside the party.
Both main parties spread fake news but the “BJP sends messages with communal and religious fundamentalist messages which Congress doesn’t. [Congress] share a lot of fake statistics, but they clearly don’t have a sound strategy,” Ms Basu adds.
Despite its huge presence in India, WhatsApp only hired its first employee in the country last year. But since then, the app has struggled to contain a torrent of false news in India — from rumours about child kidnappings to fake footage of terrorist attacks and dead bodies — that has contributed to bouts of mob violence and unrest. WhatsApp says it bans roughly 400,000 accounts in India every month.
In response to legal threats from the Indian government starting last July, which demanded that it make changes to how it operates in order to improve accountability, WhatsApp appointed a grievance officer to deal with complaints from hundreds of millions of users and hired its first-ever India head, Abhijit Bose, in March.
The biggest challenge is that, unlike Facebook, WhatsApp cannot identify the source of a message without breaking its encryption system. Instead, it has worked to make sharing more difficult, including limiting the number of recipients of a forwarded message, reducing group sizes and allowing users to decline group invitations. Other product changes being tested include a fact-checking service for images received through the app, although it would not be rolled out in time for the election, Mr Woog says.
These measures to limit virality have had limited impact, according to independent academics. “We see many instances where the same message was sent on multiple groups, over 20 groups within a 10-second window, that means there is a person, or software sending the messages,” says Mr Garimella.
WhatsApp says it has also spent about $10m in India to run a public education campaign around the dangers of misinformation on traditional media such as television, radio and newspapers. “I think I would say without hyperbole it’s probably the largest public education campaign about misinformation ever undertaken,” says Mr Woog.
The company is working with third-party organisations such as Boom Live, one of India’s independent WhatsApp fact-checkers, and AltNews, as well as the non-profit group Proto, on official fact-checking services for the duration of the election.
The efforts mirror those by non-profit First Draft during the Brazilian election, where a consortium of journalists fact-checked more than 65,000 tips and messages received from users via a dedicated WhatsApp number.
“Most of the stuff we are busting or verifying owes its origin to WhatsApp. There is fake news on Facebook too, but the numbers are small compared with WhatsApp. We get hundreds of reports each day,” says Govindraj Ethiraj, founder of BoomLive. Facebook has recently removed hundreds of “inauthentic” pages.
While Facebook and WhatsApp are confident they have ensured the smooth running of their platforms in election season, academics fear that the millions of people coming online for the first time in emerging nations via WhatsApp will be most vulnerable to this type of rumour.
“It’s great that WhatsApp . . . [is] putting out advertising campaigns, and other traditional mechanisms but my concern is that it won’t reach the people actually using it,” says Aviv Ovadya, founder of the Thoughtful Technology Project in California. “They could provide it in WhatsApp itself, make it fun, silly, interesting, in a way that people will engage with it. They should be doing a lot more.”