By: ilma hasan
June 1 2024
Did Meta violate its own policy by permitting misleading and divisive ads during Indian elections?
Share Article:
(Source: Modified by Logically Facts)
A number of Facebook pages with followers ranging from a few thousand to millions have been publishing disinformation and propaganda, targeting the Indian Muslim community and the Opposition INDIA alliance, amid the 2024 Indian general national elections.
We reviewed Meta’s ad library between February 9 to May 18 and found that some of these pages with a low follower count have spent millions on Facebook to post misleading content to increase reach. Experts, including a former member of Meta’s South Asia policy team, told Logically Facts this is a clear violation of Meta’s ad policy. Logically Facts identified these pages by going through the ad library’s top spenders data, which was predominantly anti-Opposition and pro-BJP.
A significant number of posts in the last few weeks have shared false information about the leading Opposition party Congress’ manifesto, alleging it favors Muslims — a narrative that gained traction following a speech by Indian Prime Minister Narendra Modi and other senior leaders of the ruling Bharatiya Janata Party (BJP).
Although the narrative was widely debunked, the rhetoric gained momentum and was repeatedly used by star campaigners, eventually resulting in its unbridled posting and sharing on pages amassing millions of impressions.
Pages posting misinformation pour in millions on ads
Created in December 2023, although Facebook page Political X-Ray has just about 9,300 followers, it has been one of the top spenders on Facebook ads in India amounting to Rs. 1.1 crore in the last 90 days as of May 26.
Screenshot of Meta’s ad library’s data on top spenders. (Source: Meta Ad Library/Modified by Logically Facts)
The page posts political illustrations, and all of them use misinformation and divisive narratives to target the Opposition alliance, particularly Congress’ Rahul Gandhi, Aam Aadmi Party (AAP) head and Delhi Chief Minister Arvind Kejriwal, and West Bengal Chief Minister and Trinamool Congress’ Mamata Banerjee.
Posts on Political X-Ray alleging Congress favors Muslims. (Screenshot/Modified by Logically Facts)
A post depicts an illustration of Gandhi squeezing wealth out of a saffron-clad man to give to seemingly-Muslim men. The claim is the same as the rhetoric accusing the Congress and its leaders of collecting wealth and redistributing it to minorities, particularly Muslims. The page paid for advertising between May 7 and May 9 and spent approximately Rs. 50,000 to Rs. 60,000 making over a million impressions, according to Meta’s ad library data.
Another illustration depicts reservation benefits being handed to men in skull caps, alluding to Muslims, with India’s marginalized communities the Scheduled Caste (SC), Scheduled Tribes (ST), and Other Backward Classes (OBC) being ignored. The page spent around Rs 1,500 to Rs 2,000 on this post that resulted in it making between 250,000 to 300,000 impressions.
The same narrative has been shared repeatedly through different illustrations. A reel on the page claims that the Congress party will give religion-based reservations to Muslims. The video repeats the false claim that the ruling Congress government gave reservations to Muslims in the OBC category in the southern Indian state of Karnataka. Nearly Rs 60,000 to Rs 80,000 were spent on multiple ads on this video on Facebook and Instagram, making over a million impressions.
Divisive narratives shared by Political X-Ray, and amount spend per post. (Screenshots from Meta’s ad library/Modified by Logically Facts)
Meta's ad policy stipulates that advertisers who want to make ads on elections or politics need to go through an authorization process with a “paid for by” disclaimer on ads. The social media company also mandates creators to provide functional contact information such as phone numbers and email addresses, and it bans sale, purchase or exchange of accounts.
There is contact information on Political X-Ray’s page, but no further details on who is behind the account.
Platform policy expert Devika Malik said, “The community standards that apply to organic content; also apply to ads and ads also go through an authorization process. Political ads require disclaimers around who paid for the ad and are required to comply with legal and regulatory standards prescribed in the jurisdiction that they are issued. But the initial review or pre-screen is quite elementary and we see a very high prevalence of ads that violate or don't comply with Meta's own policies as well as local law.”
Malik has previously worked in Meta's South Asia policy team on issues of hate speech, misinformation and violent extremism.
Tech Transparency Project, a research hub exploring the influence of the major technology platforms on politics and policy, noted in a May 6 report that a black market for accounts that already have approval for political advertising is thriving in India amid the ongoing elections. “It is easy to see how bad actors could use accounts like these, under fake or stolen identities, to amplify disinformation or incendiary messages,” the report concluded.
In its policy rationale, Meta states that it will remove misinformation “likely to directly contribute to the risk of imminent physical harm” and likely to contribute to interference in the functioning of political processes.
But Logically Facts’ analysis of these pages found the content uses Islamophobic narratives for political rhetoric. For instance, another Facebook page called ‘Bharat Todo Gang’, which translates to ‘break India gang’, spent over Rs. 10 lakhs on Facebook ads posting misleading and divisive content. The page has over 99,000 followers on Facebook.
“The degree of responsibility on platforms is even higher in the context of ads, where there is a direct monetary exchange with the offending actors,” Internet Freedom Foundation’s Executive Director Prateek Waghre told Logically Facts, adding that multiple investigations have shown that the improvements the platform has made to its policies are still exploitable.
Policy violations
Another page, India 272+, with over 2.8 million followers, has been posting similar content.
A post captioned, “Who is the Congress trying to benefit?”, with an image of Gandhi in a skull cap insinuates that the Muslim community will benefit more from the Congress party’s Mahalakshmi scheme, which promises cash transfer of Rs one lakh annually to a single woman from an economically weaker household. The post says, “Those with two wives, get two lakhs. Who are they conspiring to help?”
Aji Haan, with over 660,000 followers, churned similar content revolving around narratives targeting Congress’ manifesto.
Posts targeting Congress on Aji Haan and India 272+. (Source:Screenshot/Modified by Logically Facts)
While a disclaimer in the privacy and legal info section of the page states that the content is a “work of fiction” and should not be misconstrued as “genuine and true”, more scrutiny is required to determine whether such disclaimers prevent the spread of misinformation.
The page also cross-posts on other platforms, and its X (formerly Twitter) page has over 89,200 followers, including Modi.
Screenshot of AjiHaan’s X account followed by Prime Minister Narendra Modi. (Source: X)
In March 2024, ahead of the Indian general elections, Meta stated that it will remove content that could suppress voting, or contribute to imminent violence or physical harm. But a study by international civil society organization Ekō and India Civil Watch International underscored that despite commitments, advertisements propagating hate speech and disinformation were rife on the platform during elections in India.
Another report published by the watchdog groups along with The London Story on March 22 found far-right shadow advertisers spent one million dollars over 90 days on Meta, accounting for nearly 22 percent of the total sum of “issues, elections or politics” advertisements in India from February to March.
Speaking to Logically Facts, one of the researchers of the report, Maen Hammad, said, “There is this practice by many of the big spenders, that they first post the content as organic content. I could see using CrowdTangle (a soon-to-be-defunct analysis tool by Meta), many of these posts were getting viral engagement, and it was at that time it was monetized and put into an ad.”
He said that Meta is not flagging the content or the ad as problematic and the ones running the ads are putting money behind the content that performs well organically. “It's a very calculative and exploitable model. You have a network that's able to test, correct, and change content, then target voters in specific swing states to amplify these divisive narratives,” Hammad added.
The company has previously come under scrutiny for allowing political ads claiming the U.S. 2020 presidential elections were rigged. The European Commission has also launched formal proceedings against Meta for failing to monitor deceptive political advertising in the upcoming European Union elections.
Is Meta doing enough?
In response to Ekō and India Civil Watch International’s report on slander and the meme-network, Meta responded on May 13 and said that it took action against 14 accounts and admins that would need to fulfill additional identification requirements before being allowed to run ads, but on reviewing the other 38 ads flagged in the report it said that the content did not violate its advertising standards.
While the platform claims to have taken action against such accounts and admins after the report flagged content up till March this year, Logically Facts found some of the same pages like Political X-Ray continued advertising and pushing anti-Muslim and political content. A way such pages manage to circumvent regulation is via claims of it being satirical.
“Political satire is a frequent blindspot when it comes to content review. While it's hard to apply very formulaic policies in this area, it's reasonable to assess a piece of content based on harmful stereotypes, hate (which includes calls for economic and social exclusion) and also veiled threats against members of a particular community,” Malik said.
Waghre says the ads attempt to exploit conditional exemptions for satirical posts set by Meta’s policies on hate speech. “While it may not be possible for platforms policies to cover all cases, when motivated actors are actively trying to bypass their policies, these particular gaps of using 'borderline speech', memes/satirical formats, and other forms of 'malign creativity' to evade content moderation actions have been around long enough that they should be able to respond more effectively, especially in additionally risky contexts, such as Islamophobic messages in India,” he said.
The rhetoric posted on these pages garners millions of views without ever being held accountable for who runs or funds them.
Meta’s policies state that it partners with experts in determining what constitutes misinformation in crucial categories, and whether it is likely to directly contribute to the risk of imminent harm. Citing election disinformation that led to violence in Brazil in 2022, Hammad says while every country is different, platforms should constantly enhance and tailor their policies in 2024 — the year of elections.
A response from Meta is awaited. Logically Facts first reached out to Meta on May 20 and this story will be updated if and when we receive a response.
(Editor's Note: We have refrained from adding any links to the pages mentioned in the story to avoid giving them more traction.)