In the months leading up to the Jharkhand assembly election, advertisers have spent at least Rs 2.25 crore on political promotions on the Meta-owned platforms Facebook and Instagram, targeting users in the state.

The state will go to the polls in two phases, on November 13 and November 20.

Thirty-six percent of this amount was spent on advertising by third-party “shadow accounts” that promoted the Bharatiya Janata Party and included communally divisive messages, research by a group of civil-society bodies has found. Among these posts were ones that showed Muslim men as rapists and criminals, and depicted Chief Minister Hemant Soren with horns and as an insect.

Advertisement

The report based on this research notes that such advertisements are in violation of Meta’s policies, which prohibits ads that attack individuals on the basis of “protected characteristics” such as race, ethnicity, religious affiliation and caste.

“While the official page of BJP Jharkhand is using political ads to highlight issues related to government programs and electoral promises, the network of shadow accounts are engaged in posting communally divisive content and attack ads,” the report says.

It argues, “These narratives are not posted as content, but posted as ads. Meta is profiting from each of these ads posted.”

Advertisement

The report also notes that such advertisements violates India’s electoral law, including the model code of conduct, which bars parties and candidates from indulging in “any activity which may aggravate existing differences or create mutual hatred or cause tension between different castes and communities, religious or linguistic”.

Scroll emailed Meta seeking responses to the report’s findings that shadow accounts of the BJP were using the platforms to propagate communal messaging ahead of the Jharkhand election. We also emailed the BJP and the Election Commission of India, seeking responses to the report’s findings. This story will be updated if they respond.

Key narratives

The report, titled “Jharkhand’s Shadow Politics: How Meta Permits, Profits From, and Promotes Shadow Political Advertisements” was jointly published by Tech Justice Law Project, Indian American Muslim Council, India Civil Watch International, Hindus for Human Rights and Dalit Solidarity Forum.

Advertisement

It identified four key narratives that shadow accounts sought to propagate.

The first pertained to “Negative stereotypes of Muslim Men” – several posts highlighted crimes of sexual violence in Jharkhand and placed all blame for these on Muslim men and the inaction of the Hemant Soren-led government.

The second kind of narrative pertained to “Dehumanising Images of Adivasi Chief Minister”. Such posts portrayed Soren in a crude manner, such as with horns, or as an insect, while accompanying content criticised his policies and work.

The third kind of narrative pertained to “Communally Divisive Content”, which sought to pit Hindu and Muslim communities against each other. For instance, one video from the ad library, shows a group of Muslim men dressed in green and brandishing swords, chasing after a Hindu man dressed in saffron. The lone man then finds a group of Hindu men, who turn around and, armed with lathis, face the Muslim attackers approaching them.

Advertisement

The video ends with the caption, “Batenge toh katenge, sanghatit rahiye, surakshit rahiye,” or, if we are divided, we will get killed, stay united and stay safe.

The fourth kind of narrative sought to propagate the idea of a threat arising from “infiltrators” into the region. Such advertisements elaborated on the BJP’s false propaganda that Bangladeshi infiltrators from the Muslim community had caused a demographic change in the Santal Pargana region of Jharkhand, and that these infiltrators were marrying Adivasi women to gain control of legally protected Adivasi land.

Advertisements’ reach

On October 31, as people across India celebrated Diwali, between 60,000 and 70,000 Instagram users encountered an ad on the platform, posted by the “Jharkhand Chaupal” account, that linked the festival to the elections. It exhorted them to remember, as they lit their diyas, that if the current Hemant Soren-led government came to power again, Muslim infiltrators would not let them light lamps on Diwali again.

Advertisement

The ad depicted a group of angry looking Muslim men in the background with a photo of chief minister Hemant Soren in the foreground, with red horns.

Another ad published by the same account on Instagram on October 20 showed a photo of the leader of the opposition Rahul Gandhi eating a jalebi, along with an image of a pot and a glass of hadiya, or rice beer. The caption stated: “After the success of a jalebi factory in Haryana, Rahul Gandhi is coming to start a hadiya factory in Jharkhand.”

Hadiya or rice beer is an important cultural artefact of Adivasi communities in the region – but some disparagingly claim that it has led to a rise in alcoholism. There is no record of Rahul Gandhi having made any such assertions as those shown in the ad. This ad, too, was seen by between 60,000 and 70,000 users.

The authors of the report raise grave concerns about the reach these accounts enjoy. They note, in fact, that between September 22 and October 30, the single shadow account they identified that had the highest reach, Jharkhand Chaupal, spent less money on fewer ads than the BJP Jharkhand’s official account, but received tens of millions more impressions.

Advertisement

Specifically, the BJP spent Rs 97.09 lakh on 3,080 ads which received 100 million impressions, while Jharkhand Chaupal spent Rs 30.24 lakh on 1,399 ads which received 134.7 million impressions.

“Although the BJP Jharkhand account spends more than thrice of what the shadow page spends, and puts out around thrice the number of ads, it receives less impressions than Jharkhand Chaupal,” the report said. It added that “one rupee spent by the shadow network is seen by four times as many people as one rupee spent by the BJP Jharkhand page”.

Less money, more impressions

The researchers note that the figure of Rs 2.25 crore they arrived at was spent between August and October. The figure “is likely to be a conservative estimate only accounting for the ads that declare themselves to be political ads”, they explain.

Advertisement

The BJP Jharkhand’s official page accounted for 43% of this expenditure, or Rs 97.09 lakh. This money bought the party 3,080 advertisements, which garnered at least 100 million impressions.

But the authors found that shadow advertisements received a far higher number of impressions. Over their period of research, they identified at least 87 pages “actively purchasing ads on Meta linked to BJP narratives”, which accounted for 36% of the total expenditure on political advertising on Meta platforms, or Rs 81.03 lakh. This money bought these advertisers 14,559 advertisements, which garnered 450 million impressions, the report states.

Thus, the report observes, “For almost the same amount of money, shadow pages are thus putting out more content and reaching more people. This raises serious questions about Meta’s AI and differential pricing around political advertising.”

Advertisement

In contrast, the report notes, they did not find “the presence of any other party actively spending on political ads in Jharkhand” on Meta platforms.

The report also argues that though Meta claims to have a process in place to verify the identities of political advertisers, including confirming their contact information, this process has proved ineffectual in the case of these accounts.

In some cases, listed mobile numbers “appeared to be switched off when we reached out to the number on different days and times”, the authors note. In some cases, “necessary transparency requirements such as website address is missing from the profiles of these shadow accounts and the address mentioned in the profiles are generic and cannot be pinpointed on a map”.

Advertisement

Siraj Duta, an activist from the Loktantra Bachao Abhiyan, a coalition of civil bodies in Jharkhand, said, “This report again exposes the BJP’s communal agenda, its undemocratic electoral methods, and the election commission’s complicity.” He noted that the Election Commission’s failure to curb these advertisements was consistent with its reluctance to act against “BJP leaders’ repeated hate speeches and religious polarisation in the election campaign”.

Identifying shadow advertisers

In order to identify likely shadow advertisers, the authors relied on the Meta Ads Library, which allows users to access information about ads purchased on Meta platforms.

The authors created “a repository of keywords capturing the narratives and campaigns being adopted by the major political parties” ahead of the election in Jharkhand and identified accounts that used these keywords. Further, they identified the top spenders on advertisements targeting users of Meta platforms in the state during this period.

Advertisement

Next, the authors studied individual accounts from this pool and analysed the history of advertisements they had posted, as well as details pertaining to account transparency, such as owner names, website links and contact details.

Using this information, the authors arrived at a shortlist of accounts “likely to be shadow advertisers”. Through Meta’s library, the authors could access details of how much money these accounts spent, as well as the number of impressions that individual advertisements received. This allowed them to identify “a huge BJP-boosting shadow network putting out political ads on Meta, spending almost as much as the official BJP accounts”.

This, the report stated, was consistent with previous reporting by outlets such as Al Jazeera, The Reporters Collective, The Wire, Alt News and BOOM, which “have exposed the active presence of shadow advertisers on Meta since 2019”. It noted, “These reports have shown how shadow advertisers often outspend official party accounts, are full of speech that violates India’s electoral laws in terms of communal content, attack on religious minorities and incitement to violence.”

Among the report’s recommendations to Meta are that it “improve political advertiser verification”, take action against advertisers that “repeatedly run ads in violation of Meta’s guidelines, containing hate speech and operating in a coordinated manner” and improve the depth of information available on the platforms about political advertisers and their content. As long as such advertisers circulate, the authors observe, “The real loser in all of this is the Indian voter, subject to a pervasive, poisonous online discourse.”