“Stupefying speed.” Bloomberg checked in with Vishvas News, Facebook’s largest Indian-language fact-checking contractor, to see how things went during India’s general elections. (Results announced Thursday gave Prime Minister Narendra Modi’s BJP party a landslide victory.). Pallavi Mishra, Vishvas’ manager, “spent two weeks recently talking with internet users in small cities. She found most people are so new to social media they have no clue about bogus content. They share stories indiscriminately, with stupefying speed. ‘Being the “first” to share things in their circles gave them a rush,’ she says.” From the story:

About 90% of internet users coming online today are non-English speakers, often with more trusting attitudes than experienced surfers. In one Hindi example, a phony Facebook post for job vacancies at the subway train company in the city of Jaipur asked those interested to send personal details. Within hours, about 20,000 had responded with names, emails and phone numbers.

India has 900 million registered voters, an estimated 67 percent of whom voted, but Vishvas News has just seven full-time fact-checkers. India has 23 official languages; “Facebook has hired contractors to verify content in 10 of those languages, but those staffers are spread thin and posts in more than a dozen other languages — Sindhi, Odia and Kannada among them — are completely unvetted.”

A February 2019 study by Microsoft found that, out of 22 countries, Indian citizens are the most likely to encounter fake news and hoaxes. And “the proportion of polarizing political news and information in circulation over social media in India is worse than all of the other country case studies we have analyzed,” wrote researchers from Oxford’s Project on Computational Propaganda earlier this month, “except the US Presidential election in 2016.”

Here’s more from that report:

We find that (1) more than a quarter of the content shared by the Bharatiya Janata Party (BJP) and a fifth of the content shared by the Indian National Congress (INC) is junk news, while the Samajwadi and Bahujan Samaj Party (SP-BSP) shares very little sensational, extremist, or conspiratorial content. (2) For visual content being shared in our sample of WhatsApp groups, a third of the BJP’s images, a quarter of the INC’s images, and a tenth the SP-BSP’s images were catalogued as divisive and conspiratorial. Comparing the platforms, we find that (3) misinformation on WhatsApp primarily takes the form of visual content, while misinformation on Facebook involves links to sensational, extremist, and conspiratorial news sites and visual content

Current estimates have the junk-spreading BJP and INC winning 297 and 54 seats in India’s lower house of parliament, respectively, while the comparatively junk-free SP-BSP gets 19.

“Bigger than all far-right party pages combined.” The European Union is also heading into elections this week. Global activist network Avaaz released a report this week outlining a three-month investigation that turned up more than 500 suspect EU-related Facebook pages and groups, “which were followed by nearly 32 million people and generated over 67 million interactions (comments, likes, shares) in the last three months alone.” Avaaz reported its findings to Facebook and says that “77% of the pages and groups reported, accounting for about 20% of all interactions across the reported networks,” have been taken down.

From Wired’s Issie Lapowsky:

In Germany, for example, Avaaz says Facebook took action against 131 fake accounts, many of which propped up the far-right party Alternative for Germany, or AfD. In France, Facebook removed a page called Suavelos, which promoted white-nationalist ideas, something Facebook only recently prohibited. Facebook shut down 23 pages Avaaz reported in Italy for a variety of infractions, including changing the name of a page to something unrelated to the page’s origins; one such page changed its name eight times, according to Avaaz, starting as a sports page and eventually morphing into a political page. In Poland, Avaaz found a network of pages sharing a fake news story almost simultaneously about migrant drivers raping European women. Of the 26 pages Avaaz found that shared the article, Facebook took down 11. Avaaz was also behind a recently reported takedown of pages and groups in Spain just before the election there in April. Finally, in the UK, Avaaz says Facebook removed 132 posts, pages, and groups listed in its report.

These takedowns are different from the ones Facebook has traditionally announced publicly, in that none of them appear to be what the company calls “coordinated inauthentic behavior” driven by a nation-state or single organization. Facebook has previously announced these types of influence operations associated with entities in Russia, Iran, and Israel. Instead, the Facebook spokesperson said, these removals are part of actions the company takes “regularly and routinely” against content that violates its policies.

Other European studies had similar findings. My colleague Christine Schmidt wrote up one this week. The Financial Times wrote up another:

Researchers found many new websites, notably in Italian, set up to collate and disseminate anti-migrant and anti-Muslim stories. In some cases, the stories are linked to current events such as the fire at Notre-Dame cathedral in Paris. 

“Anti-migrant themes have been relentlessly exploited as a way to galvanize, connect and create active digital communities,” said Alex Romero, chief executive at Alto Data Analytics, which found anti-immigration content was among the most-shared material on more than 500 Twitter accounts suspected of spreading disinformation or hyperpartisan stories. 

“It’s not 2016. We are not seeing the automated, networked activity with an obvious Russian fingerprint across these elections,” Sasha Havlicek, head of London’s Institute for Strategic Dialogue, told The Washington Post. “What we’re seeing much more of is coordinated, transnational far-right information operations.”

A 24-hour lifespan for real news? In a presentation for a recent conference on information management, Anjan Pal and Alton Y. K. Chua of Singapore’s Nanyang Technological University looked at how fake news versus real news stories spread differently on Twitter. One key difference is the amount of time a story continues to be shared after first publication:

The tweet volume of real news dropped drastically after the first 24 hours (e.g., real news #1 had 37 tweets after 24 hours, and 38 tweets after 96 hours). This was not true however for fake news (e.g., fake news #1 had 226 tweets after 24 hours, and 452 tweets after 96 hours).

Results indicated that the tweet volume of real news dropped drastically after the first day. However, this was not true for fake news. While real news seems to lose its value after a day, fake news sustains its popularity.

This was a really small study — Pal and Chua looked at just 20 fake news and 20 real news stories, all about Hillary Clinton — but the finding about volume after the first day echoes findings from other studies. A large 2018 MIT study, for instance, also found that fake news spread more broadly and more quickly than real news. A 2016 study showed that the first 24 hours are key for a news story’s spread on Twitter, and a study of legitimate news stories in 2012 found that retweeting “typically [ended] between 10 and 72 hours after an article [was] originally shared.”