Maryam Shafipour spent two years in Tehran’s notorious Evin Prison, more than three months of it in solitary confinement, as the Iranian regime tried to break her. Years later, she’s in exile in Canada, and the regime, or its supporter base, appears to be trying a different tack: a steady barrage of online abuse and innuendo. On Instagram, Telegram, Twitter and WhatsApp, a quick search of Shafipour’s name in Farsi finds dozens of posts featuring a video that she says sketches out lies about her finances and sexual activities. On Instagram, someone paid to promote one of the posts like an ad, showing the video to more people. Worse, accounts that sort of look like they belong to Iranian progressives regularly imply that she’s been compromised and switched sides.

The aim is “to isolate me from my community,” Shafipour says. “They don’t do it for just a short period of time. They do it for years. They continue telling lies about you, and it really works, after years.” She has struggled with depression and post-traumatic stress disorder, linked in part to time she spent in prison starting in 2014 on charges of “propaganda against the regime” and “assembly and collusion against national security.” While she was locked up, Shafipour’s jailers tormented her by lying about her mother having died. When her mother really did die, in December 2019, the online libel grew worse: Videos began to circulate calling her a spy for the regime.

Facing challenges to its power and international criticism over its human-rights abuses, the Iranian government has built a large and sophisticated cyber army, which has targeted its domestic and foreign opponents using cyberattacks, propaganda and large-scale misinformation campaigns, according to researchers and activists who monitor online activity in Iran. The Islamic Revolutionary Guard Corps (IRGC) and other forces have used networks of fake news websites and social media accounts to seed stories in multiple languages. “It’s a kind of war,” Shafipour says. “An unseen war.” She and her peers say they’ve noticed these operations intensify since late last year, after tens of thousands of Iranians took to the streets to protest the death of 22-year-old Mahsa Amini in the custody of the country’s “morality police.”

As the brutality of the government response to the protests has reverberated across social media, activists (including Shafipour) have asked tech companies to do more to limit the regime’s reach. For starters, they say, the companies should ban the regime’s key figures from their platforms. This includes Supreme Leader Ayatollah Khamenei, who has millions of followers on Instagram and Twitter, the latter of which is blocked in Iran.

So far, the activists have made little progress. They, like others who’ve tried to compel tech companies to take action against dangerous accounts, have found that the companies are shielded by Section 230 of the US Communications Decency Act of 1996, which says they aren’t considered publishers and therefore aren’t liable for their posts. The 26 most important words in tech, as the provision has been called, are being challenged at the US Supreme Court. In late February, the court heard oral arguments in Gonzalez v. Google LLC, brought by the family of an American student killed in a terrorist attack in Paris. The perpetrator was allegedly radicalized by content that YouTube’s algorithm recommended for him. (Twitter v. Taameneh, a second pending case brought by surviving family members of a terrorist attack victim, alleges that Facebook, Google and Twitter all bear responsibility for radicalizing users.)

In court, and in a subsequent statement, Google argued that Section 230 protects free expression and is “the economic backbone of the internet.” Meta Platforms Inc. and Twitter Inc. made the same case in briefs filed to the court in support of Google in the Gonzalez case. During oral arguments in the Taameneh case, Twitter’s counsel denied responsibility, arguing that allowing Islamic State on its platform doesn’t constitute knowing assistance to the militant al-Qaeda splinter group.

Analysts say the high court isn’t likely to strike down the law anytime soon. Tech companies, and many digital-rights groups, have defended Section 230 as a fundamental underpinning of freedom online. But activists and some digital-rights advocates stress that the companies don’t need to be compelled by the force of law to decide whether they ought to amplify the messages of a regime that’s executing protesters and arresting people for their social media posts. “It’s crazy for me that state leaders that don’t allow their own citizens to use the platforms are allowed on the platform,” says Marie Lamensch, a project coordinator at the Montreal Institute for Genocide & Human Rights Studies at Concordia University who researches online hate speech and violent extremism. “They shouldn’t be.”

Shafipour was a student activist at Imam Khomeini International University in the late 2000s, where she campaigned for reforms including women’s rights. In 2009 she joined the presidential election campaign of the reformist politician Mehdi Karroubi. After irregularities in that election led to huge protests, dozens of people were killed and thousands arrested in the regime’s crackdowns. Online, the activists made effective use of emerging forms of social media to organize and to broadcast the protests and the regime’s violence to the world. So the authorities tried to take control of cyberspace, too.

Iran’s government pioneered tactics that would become standard for authoritarian regimes. It shut down the internet for extended periods of time. It began blocking the social media platforms where Iranians congregated (at the time, Facebook and Twitter). But the regime also started to develop more subtle tools, honing technical skills and refining infrastructure to monitor and manipulate those social networks. While censoring Facebook and Twitter in Iran, the government has made use of the platforms overseas to try to shift public opinion.

Iran has been subject to international sanctions since 2006, after it refused to comply with a United Nations resolution calling for it to halt its nuclear weapons program. The US and its allies have accused the country of human-rights abuses and supporting terrorist groups. Online propaganda campaigns give the regime a chance to downplay such allegations, says Abde Amr, a researcher at the Atlantic Council think tank’s Digital Forensic Research Lab.

Iran’s overseas influence operations are broadly similar to some of the so-called hybrid-conflict tactics that Russia has pioneered, albeit less extensive. In 2018 the cybersecurity company FireEye identified a network of websites it linked to the ayatollah’s forces that claimed to be progressive US and UK news outlets with names such as Liberty Front Press and the British Left. These sites published articles criticizing US and UK involvement in the Middle East and the role of American allies, particularly Saudi Arabia, in conflicts in Yemen and elsewhere in the region. Other sites amplified stories about racial and social injustice in the US and the UK.

Linked accounts, many of them fake, spread this material on social media. In 2018 and 2019, Facebook removed about 1,000 pages and accounts from Facebook and Instagram for engaging in “coordinated inauthentic behavior tied to Iran.” Around the same time, Twitter removed roughly 3,000 accounts that it says might have originated in Iran.

For Iranians who get around website restrictions using virtual private networks or who live abroad, the state’s use of social media has served as an extension of its existing media and propaganda outlets. Researchers say regime-linked accounts are still spreading misinformation on Twitter in English and Farsi. Accounts on Instagram and Telegram now routinely share videos of what appear to be forced confessions or recantations by people involved in protests, apparently aimed at emphasizing the strength of the regime and the futility of resistance. Some are directly linked to state media, and others are anonymous, but researchers say they believe these accounts are generally operated by the state or its proxies. “State media has been used as, basically, an organ of torture and repression,” says Mahsa Alimardani, an Iranian researcher at the Oxford Internet Institute. “The digitization of the propaganda systems for the Islamic Republic fed into what they were doing already.” Other accounts, thought to be operated by the intelligence services, impersonate progressives, royalists or members of opposition groups and post material designed to discredit them or otherwise sow division, she says.

Whenever social movements flare up on the Iranian internet, they are quickly targeted. In February and March of 2022, a group of women’s rights activists took to Instagram to share their experiences of sexual violence, inspired by the global #MeToo movement. In April, their accounts were suddenly flooded with thousands of new followers, which appeared to be bots. Qurium Media Foundation, a digital forensics nonprofit, traced the fake followers back to a pay-to-play bot farm in Pakistan. The motivation for the campaign wasn’t entirely clear, but it could have been an attempt to get the women’s accounts suspended. (Instagram often suspends accounts with large numbers of fake followers.) It could also have been simply a tactic of intimidation.

Since fresh protests began in Iran in September, the influence operations have gone into overdrive. Telegram channels, which activists and researchers link to the intelligence agencies and the IRGC, have been crowdsourcing the identities of protesters and broadcasting images and videos of brutal crackdowns and what appear to be forced confessions. Instagram accounts have tried to hijack anti-regime hashtags with pro-regime messaging.

The @khamenei_ir account, which tweets the thoughts of Iran’s supreme leader in English to 952,000 followers on Twitter, took 17 days to acknowledge the latest unrest in the country. In the weeks after Amini’s death, the account mused on the causes of the Iran-Iraq War in the 1980s, discussed relations with China and ruminated on the virtue of patience. Finally, on Oct. 3, it addressed the violence, using the #MahsaAmini hashtag that supporters of the protests had set up. Khamenei blamed the uprising on “schemes designed by the US; the usurping, fake Zionist regime; their mercenaries; & some treasonous Iranians abroad who helped them.”

Twitter and Meta have sporadically taken down content and suspended regime-linked accounts, often after users have reported them for violations of the platforms’ terms of use. Both companies are known to apply their policies differently for high-profile political leaders, typically giving them more leeway to break the rules. (Then-US President Donald Trump’s Facebook and Twitter accounts were suspended in January 2021 following the attack on the Capitol by his supporters, but both have since been reinstated.) Critics say there are strong moral and practical arguments against doing this. Faced with social and political pressure, tech companies have invested in trying to reduce the amount of “inauthentic activity,” hate speech and other dangerous posts on their platforms. This effort has often meant a focus on bots, fake accounts and less overt espionage. But official accounts play an important role in the disinformation ecosystem. The overt and covert components of information operations often overlap and intersect, with narratives seeded on social media eventually amplified and legitimized on state-owned media.

One attempt to force the major platforms to shut down the Iranian regime’s official accounts has already failed. It was initiated by Rumi Parsa, an Iranian-born corporate lawyer living in California, who says he was outraged to find the ayatollah and members of the IRGC on Instagram and Twitter. Parsa approached Twitter and Meta—which owns Instagram, where Khamenei has more than 5 million followers across several accounts—to ask them to consider banning the ayatollah and related accounts. “I actually begged,” he says. “People are getting killed.”

When the platforms declined to take action, he sued. The case, filed in California, claimed that by letting Khamenei and other Iranian officials use their platforms, the social media companies were providing services to the regime in breach of sanctions. It also claimed that they had breached their own terms of use by allowing the ayatollah to post antisemitic hate speech and calls to violence. The @khamenei_ir account—and others in Arabic, French and Spanish—has previously presented statements that can be interpreted as direct threats against Israel and the US and as attacks on Jewish people. The account has also been accused of inciting last August’s assassination attempt against Salman Rushdie in Chautauqua, New York. In 2019 it reaffirmed the 1989 fatwa by the previous supreme leader, Ayatollah Khomeini, which ordered Rushdie’s execution after the publication of his novel The Satanic Verses. Twitter briefly limited the account’s ability to post after that tweet, but Khamenei remained on the platform.

In January, Parsa had to withdraw the suit. He failed to establish standing in federal court, meaning he couldn’t prove he had sufficient personal connection to the alleged harms. Along with Section 230, this is a common challenge for people taking American tech companies to court. The bar for proving harm is very high.

Without help from the platforms, activists have had to play a kind of whack-a-mole against individual accounts or seek changes to policies around more specific kinds of material. They’ve won small victories. In January, after two years of campaigning, they managed to persuade Facebook’s oversight board, an independent body that examines moderation decisions, to recommend overturning a ban on the phrase “marg bar Khamenei”—a common Farsi expression in protest posts that translates literally as “death to Khamenei” but actually means “down with Khamenei.” Later that month, they persuaded Instagram to block a counterprotest hashtag they say has been used to spread disinformation. (It remains on Twitter.)

Shafipour says activists are growing more adept at identifying and countering the regime’s attacks online. “We both hurt each other,” she says. They are winning small victories, getting individual things suspended or policies changed, and there are still protests happening online and offline, in Iran and in foreign capitals. The regime hasn’t won in cyberspace yet. But by continuing to amplify its messages, she says, tech companies are keeping the fight that much more asymmetrical.

Things on Shafipour’s side are getting tougher, too, she says. Since Elon Musk took over Twitter last year, the teams in charge of moderation and safety have been decimated, and it’s now much tougher to get the company to suspend accounts that spread lies or post videos of forced confessions. Twitter has even made it easier to spread misinformation, she says: By selling the blue check marks that signal verification, it has allowed bad actors to buy the appearance of authenticity.

“A lot of cyber army [accounts] from the regime have a blue mark now,” Shafipour says. “It can really change the game against us.”

Source » bloomberg