Arabic social media remains an unchecked Wild West

Suspected Kremlin agents peddle falsehoods masquerading as Instagram models. The terrorist organization Hezbollah posts propaganda updates as if it were a news organization. More than 2 million Iraqis join Facebook groups where guns are bought and sold without checks.

Welcome to the world of Arabic-language social media — a Wild West where content moderation is minimal, foreign governments act with abandon, and jihadists foster online hate in arguably some of the world’s most war-torn countries.

Facebook, Twitter and other platforms have championed their efforts to clamp down on misinformation, politically divisive material and outright bigotry across the United States and European Union. But such checks — and any ensuing enforcement of these companies’ rules — have often been absent in Arabic-language social media, based on analysis by Digital Bridge, POLITICO’s transatlantic tech newsletter.

The failure of tech giants to sufficiently monitor what happens in Arabic — one of the fastest-growing demographics on social media, based on yearly new users, with 350 million Arabic speakers worldwide — highlights the disparities between how these companies are trying to crack down on harmful content in the West, while still missing scores of similar material in places like Syria, Afghanistan and Iraq. Many of these countries remain gridlocked by sectarian tension and where online hate can quickly spill over into offline violence.

For disinformation analysts, campaigning groups and some politicians within the region, it feels like a two-tier system, where one set of rules applies to Europe and the U.S., and another applies to the rest of the world.

“Groups don’t have to hide what they are doing because the moderation is so full of holes,” said Moustafa Ayad, executive director for Africa, the Middle East and Asia at the Institute for Strategic Dialogue, a think tank that tracks online extremism. Ayad worked with POLITICO to uncover how a large amount of Arabic-language harmful content and state-backed disinformation remain widespread online.

“It’s just so frustrating,” he added. “[The social media companies] did investigations around the U.S. elections, even though they were late. It doesn’t feel like they’re doing that in [the] Middle East, proactively or reactively.”

In response, Meta, the parent company of Facebook and Instagram, acknowledged that it still had more work to do when combating online hate in Arabic, but said it had increased the number of regional experts and made other investments over the past two years to police such material. Twitter did not respond to a request for comment.

State-backed falsehoods

Elena Kosogorov has a big following online.

Her Twitter biography claims that she is an editor within the Arabic unit of Sputnik, the Kremlin-backed media outlet. Since late February — when Russia invaded Ukraine — Kosogorov amassed more than 140,000 followers on the social media platform. She regularly pumped out pro-Vladimir Putin falsehoods about the war, disinformation decrying NATO aggression in Eastern Europe and, increasingly, anti-LGBTQ+ narratives that were shared widely within her largely male online audience.

Earlier this week, Twitter removed two of her accounts for breaking its terms of service. But within days, another account — complete with the same image and biography — was back up and again amassing a sizable Arabic-language audience. It quickly gained a renewed following after Kosogorov posted links to the account in her separate Arabic-language Telegram channel which also included a graphic video of the recent mass shooting in Buffalo, New York.

“This is the [sic] devoid of human values in the Western world: Gays are the first to receive the monkeypox vaccine, and no one else will be allowed to get the vaccine until after the gay shots have been completed,” she posted on Twitter soon after restarting her new account. “Where is the western world?” Twitter subsequently suspended the account after POLITICO flagged the content.

Kosogorov is not alone. The Institute for Strategic Dialogue’s Ayad discovered a network of at least 11 Twitter handles, some of which had stolen headshots from Instagram models, that repeatedly promoted Russian disinformation to a large audience across the Arabic world. Several of these accounts linked themselves directly to Russian state media, though POLITICO was not able to confirm that attribution.

The social media accounts, five of which were subsequently suspended by Twitter, including the three associated with Kosogorov, followed a similar game plan. They mixed pro-Kremlin images of military parades and those of Putin with photos of themselves posing seductively or wearing military uniforms. Such posts routinely garnered hundreds of shares and were exclusively supportive of Russia’s geopolitical aims while attacking perceived Western abuses.

“They’re primarily concerned with building up or sustaining a level of disdain for the West,” said Ayad, adding that Syrian- and Iranian-backed media outlets also continued to promote pro-Russian narratives in Arabic with little or no content moderation. An Iraqi news organization with ties to Iranian militias, for instance, was able to buy Facebook ads that peddled pro-Kremlin falsehoods about Volodymyr Zelenskyy, Ukraine’s president. It was subsequently deleted when POLITICO flagged the paid-for content to Meta.

War-torn countries

When Frances Haugen, a Facebook whistleblower, disclosed reams of internal company documents, it painted a bleak picture of how the company polices Arabic-language content. Meta said it had many undisclosed investments to expand its content moderation. But in a Facebook memo in late 2020, engineers acknowledged that 60 percent of Arabic-language content on the platform went undetected, and on its local content moderation team, “Iraqi representation was close to non-existent.”

Those problems have not gone away.

In four Iraq-focused Facebook groups, with a combined following of more than 2 million people, users buy and sell arms in direct violation of Meta’s terms of service, based on Ayad’s analysis. The company’s rules say private individuals cannot sell firearms on the network, and such posts will be removed as soon as Meta becomes aware of them. In the wake of repeated mass shootings in the U.S., the company has cracked down on these types of activities in its home market, but in Iraq, such Facebook-based purchases remain common.

Some of these Iraqi groups have names like “Your Weapon is Your Prestige,” and state categorically that people should not use the groups to sell guns. But in scores of comments on these groups, POLITICO discovered people routinely posting ads for weapons like AK-47s, pistols and other weapons like grenades in a country that is still plagued by sectarian conflict between Sunni and Shia militias. Last month, for instance, two professors in Iraq’s autonomous Kurdish region were killed after a disgruntled former student bought a weapon on Facebook — though not from the four groups reviewed by POLITICO.

After POLITICO contacted Meta about these groups, the company removed two of them for violating its terms of service.

For Haugen, the Facebook whistleblower, the non-English parts of the social-networking giant have not garnered the same level of attention compared with how content is moderated within the U.S. That has left hundreds of millions of people vulnerable to online attacks, hate speech and other forms of violence in places where Meta now dominates people’s online experiences.

“Facebook in English in the United States is a fundamentally different product than what is available in every other country in the world,” she told POLITICO when asked how differently the platform’s content rules were implemented, globally, compared with within the U.S. “Facebook moved into the most fragile parts of the world and became the internet.”