search

With Israel and Hamas @war, algorithms and fake accounts fueled #hatred online

Researchers expose fake accounts and extreme content that fanned the flames of outrage, fear and hatred in Israel and abroad

Simona Weinglass is an investigative reporter at The Times of Israel.

Suspected fake media accounts by foreign actors impersonating Israelis that were called out by Israel's Fake Reporter organization (Courtesy)
Suspected fake media accounts by foreign actors impersonating Israelis that were called out by Israel's Fake Reporter organization (Courtesy)

In the 11 days in May during which Hamas launched rockets at Israel and Israel responded by bombing targets in the Gaza Strip, social media itself turned into a battlefield.

Around the world, social media influencers were successfully spreading the narrative that Israelis are “oppressors” and Palestinians “oppressed,” The New York Times reported last Tuesday, three days before a ceasefire went into effect. The narrative appeared to be gaining ground, the paper noted.

Social media and messaging apps reportedly also played a role in helping far-right Jewish extremists in Israel organize violent demonstrations against Arabs, and aided the spread of inciteful rumors that may have fueled Arab attacks on Jews and vice versa.

But while support for Palestinians on streets worldwide may have been real, Israeli researchers attempting to pierce the fog of war surrounding the conflict have raised doubts regarding how much of the supposed social media war was truly authentic. Experts say that we often don’t know why we’re exposed to a particular piece of content online, whether as a byproduct of an algorithm seeking to shock and outrage us or by bad actors hoping to do the same.

“The social media networks made it very easy for anyone to tweak the system,” said Elad Ratson, a former Israeli diplomat who now runs Vayehee, a Paris-based company that uses technology to counter misinformation online.

An Instagram meme that was reportedly shared by fashion model Bella Hadid to her 42 million Instragram followers (Screenshot)

Puppets with poor Hebrew

Ratson said that while many social media users are behaving in ways that feel authentic to them, they are all constantly being manipulated by interested parties in ways they may not understand.

During the conflict, state and quasi-state actors were working covertly to steer online discussions surrounding the fighting between Israel and Hamas, he asserted.

“There are a lot of elements that finance the promotion of anti-Israeli content,” he said. “It’s actually more common than you think.”

Social media networks make it possible, he said, with varying degrees of ease, for a third party to pay for the promotion of content posted by someone they may not even know. Nor do social media companies require the identity of that third party to be disclosed.

Thus, pro-Israel actors might spend money to promote the tweet of a woman in Iowa who expressed views favorable to Israel but who had no relationship whatsoever to the country. At the same time pro-Palestinian operatives might pay to promote a social media post by a neighbor of hers who has no relationship to Gaza. Their promoted content, as a consequence, would go unexpectedly viral. The posts are authentic, said Ratson, but the only reason you would see them is because somebody paid to promote them, but you would not know who had done so.

Twitter did not respond on the record to the question of whether such paid promotions take place on its platform.

“Several governments I worked with or that have been the subject of my research, employ people who scout the social media landscape to identify narratives that they have, as a country, an interest in promoting,” Ratson said. “So for example, they will go on Twitter and identify voices that correspond with their worldview.”

More insidious, though, Ratson said, are covert accounts known as bots, sock puppets or fake-identity accounts whose purpose is to cause certain narratives to trend over others. This can be done through techniques such as “astroturfing, follower boosting or artificial hyper engagement,” he said.

Shortly after the outbreak of hostilities between Israel and Hamas on May 10, Ratson discovered a network of several hundred Twitter accounts created in recent days that were tweeting in laughably incorrect Hebrew, likely the product of Google Translate, he said.

Most of these accounts were created on the same day, and followed each other,  commented and retweeted each other’s messages, including the language errors, strongly suggesting that they were not authentic users but rather part of an orchestrated disinformation campaign.

One account, supposedly belonging to a woman named Esther from Tel Aviv, tweeted in Hebrew “my cousin works in the border guard. He says that there are many dead and the situation is very painful and scary. He said that Israel is keeping this information secret and preventing the world from knowing. Israel has been defeated!”

Tweets by a suspected fake Twitter account under the name of an Israeli woman named Esther, telling Israelis that there are many casualties and they should flee the country (Screenshot)

Another account, belonging to someone supposedly named Alex Nahamin, suggested that Israelis “go back to Germany.”

“When they told my grandmother and grandfather in Germany to come to this country they promised them peace, freedom and love from every part of the world and they gave them millions of dollars and provided them with houses and land, but what was achieved was the total opposite. We did not find peace, freedom or love from the world. [We should] go back to Germany.”

A suspected fake Twitter account belonging to a supposed Israeli named Alex Nahamin suggesting that Israelis “go back to Germany” (Screen capture: Twitter)

Another account, supposedly from a woman named Perah, retweeted, “Israel is definitely defeated. Its losses are great. But don’t show this to the media. If you want to survive, you must leave occupied Palestine forever. Things are going to develop a lot. The Israeli dead will increase.”

A suspected fake Twitter account, of a supposed Israeli woman named Perach that warns Israelis that if they want to survive they should “leave occupied Palestine forever” (Screenshot)

Ratson reported several of these accounts to Twitter.

The same sock puppet network was also flagged by Achiya Schatz, a spokesperson for Fake Reporter, a network of Israeli activists that exposes inauthentic or extremist behavior on social media.

“It’s a foreign influence campaign whose purpose is to demoralize Israelis,” he said.

While Schatz guessed the network was Iranian — Facebook, Twitter and others have accused Tehran of being linked to networks of faked accounts in the past — Ratson believes the relative lack of sophistication points to Hamas operatives being behind the operation.

“We can’t say for sure who is behind it. It might be Iran, it might be Hamas. Either way, Iran and Hamas work in close collaboration,” said Schatz.

During May’s flare-up of violence, Fake Reporter set about exposing inauthentic social media activity and misleading information being shared across messaging apps, posting information on its website and Twitter account, and asking for tips from the public.

Twitter responds

The Times of Israel sent the handles of about 140 accounts identified by Ratson to Twitter. Most of the accounts were promptly suspended and/or removed. Twitter told The Times of Israel it had removed the accounts because they were “fake,” but rejected the claim that they were part of a coordinated campaign or that there might be state actors behind them.

“Our analysis has not identified any state-backed activity or coordinated disinformation campaign at work here at this time. Rather we’ve found – and suspended – a number of fake, inauthentic accounts but nothing that is directly attributable to a so-called information operation, network, or state-backed campaign,”  a Twitter spokeswoman told The Times of Israel.

The spokeswoman declined to share information about who might be behind the fake accounts, what country they were from or how many users were exposed to their tweets.

A Twitter representative explained to The Times of Israel that only Twitter has the ability to identify coordinated disinformation campaigns by both state and non-state actors because only Twitter possesses the data necessary to make such attributions. Twitter does not share this data with the public. The representative said that independent researchers cannot reliably identify such campaigns.

The Times of Israel asked the Twitter representative if there is any outside body that can check and vet whether something is a coordinated campaign or a state-sponsored campaign or whether Twitter is the sole arbiter of that.

The Twitter representative replied: “Our disclosures are peer-reviewed and validated by independent academic entities, including partnerships with Stanford and ASPI. The data we use to determine whether or not information operations are occurring on Twitter is non-public. Our privacy obligations require them to remain non-public. As a result, we’re the only company in the industry that discloses every single public account, piece of media, and Tweet we can reliably link to a state-backed operation. This transparency approach empowers third parties to review it, hold us to account, and to further public understanding.”

The Twitter representative added that when and if Twitter identifies state-linked disinformation campaigns, “we will disclose it as a matter of routine in our Twitter Transparency Center. We do so to educate people about the tactics of bad actors on our service and to empower independent analysis of our work.”

Sock puppets in Israel

Achiya Schatz (YouTube screenshot)

The Israeli government’s alleged use of sock puppets to push its messaging online was seen as controversial inside the halls of the government, said Ratson.

In 2020, the Strategic Affairs Ministry, which had moved from the hands of the ruling Likud party to those of the centrist Blue and White party, indicated it did not endorse the use of such practices, following rumors that it had created fake accounts to influence social media in the past.

Likud, headed by Prime Minister Benjamin Netanyahu, was accused in 2019 of employing hundreds of fake accounts in the run-up to elections — which the party denied by parading some of the supposedly fake users for the press.

Fake Reporter itself was initially established by anti-corruption activists who have been demonstrating outside Netanyahu’s official residence and private home over the past year, to combat what Schatz described as disinformation campaigns attempting to discredit their movement, he said.

“We encountered tons of fake news and fake media campaigns. Some of them were by Yair Netanyahu and people around [Prime Minister] Netanyahu. But others were anonymous,” Schatz recalled. “Sometimes pro-Netanyahu activists would enter anti-corruption activists’ groups and say they planned to attend the protest the next day even though they were COVID positive. It was a lie.

“Another time someone wrote in the group, ‘I did not get money to attend the protest today.’ The idea was to suggest that he was being paid. I think the audience for this was Netanyahu’s base, to suggest that the protesters were phony,” said Schatz.

Earlier this month, Fake Reporter exposed what it described as Telegram and WhatsApp groups where far-right Jewish activists were organizing to go into mixed Arab and Jewish towns and beat up Arabs. They also posted screenshots of calls by right-wing activists to physically attack journalists.

At least 20 Israeli journalists were assaulted while covering the Arab-Jewish violence, and more were threatened online, many of them by far-right activists, leading at least one news organization to order security details for correspondents.

Anti-Netanyahu protesters wave flags, hold signs and chant slogans during a demonstration outside the Prime Minister’s Residence in Jerusalem, September 20, 2020. (AP Photo/Sebastian Scheiner)

Schatz said that the extremist language and images that his group encounters on Hebrew social media largely come from foreign actors or from the Israeli right. He said he had not encountered any social media extremism from the left side of the political spectrum within Israel, though he was troubled by reports of leftists outside of Israel saying Israel has no right to exist.

“That is very sad,” he said. “It is one of the reasons we are fighting for [the character of] this place.”

Schatz blamed ultra-nationalists inside Israel for inspiring rancor against the country from abroad.

“This delegitimizes Israel and causes hatred outside Israel and inside Israel. It’s a price we pay,” he said. “But what we can try to do is combat extremism from within Israel.”

An extremist social network?

On May 10, as Jerusalem tensions reached a boiling point just ahead of the 11-day war with Gaza and the spurts of mob violence inside Israel, a new social network was launched in Hebrew, meant to cater to members of the far-right.

Mykey.co.il was started by Boaz Golan, a minority owner of the nationalist 0404.co.il news site and a popular host on Israel’s Channel 20, a right-wing news channel.

Mykey bills itself [Hebrew] as a social network for the “hundreds of thousands of Israelis who suffer from having their content blocked on other social networks. They are silenced over nothing even though they did not break any law and did not libel anyone but merely because their worldview does not jibe with the worldview of whomever, or because a group of internet users decided to report them.”

Some of the content that could be found on Mykey.co.il during its first weeks of operation included memes that indeed would likely be removed from other social networks for violating their hate speech standards. For instance, one of the site’s administrators posted a meme that said “Kahane was right,” referring to late rabbi and ultra-nationalist MK Meir Kahane, who espoused virulent anti-Arab views.

Adiel Levy, an employee of 0404.co.il and an administrator of the new social network Mykey.co.il posted in mid-May that “Kahane was right” (Screenshot)

Another post, by a writer for the 0404 website, stated that “there is no coexistence with cancer,” using the disease to refer to Israel’s Arab population.

Other public posts called on Jews to boycott all Arab businesses within Israel; several suggested that Israel “wipe Gaza off the map.”

A writer for the 0404 website posted “There is no coexitence with cancer” on the Mykey social network in mid-May (Screenshot)

0404.co.il is a pro-Netanyahu news site with close to one million unique visitors a month, according to the website analytics site Similarweb.com

Eighty percent of 0404.co.il is owned by Vyatcheslav (Yitzhak) Mirilashvili who was also one of the founders of VKontakte.ru, Russia’s largest social network, whose former CEO Pavel Durov has been accused of secretly providing data to Russian intelligence services. Durov has denied the allegations.

In 2016, the Kremlin-linked Internet Research Agency ran covert social media campaigns ahead of the US general election that scholars believe sought to, among other goals, generate greater tribalism and polarization in American society.

A post on Mykey.co.il which claims that Jewish activists tied up a young Arab man with duct tape after catching him setting Jewish cars on fire (Screenshot)

The role of algorithms

Compounding the issue are algorithms designed by the social media platforms themselves, which influence the character of conversations online and may wind up helping the spread of inciteful content.

“All social media platforms have an interest in keeping us engaged longer with their platforms, as it increases their ability to serve us with more paid content,” he said.

“All social media platforms developed algorithms that are designed to identify the type of content we like and serve us more of that content,” Ratson said. “What is most unfortunate is that the type of content that seems to never fail in grabbing people’s attention is the controversial, the gory, the gruesome and the violent.”

While all major social media platforms have terms of service that prohibit the posting of violence, in Ratson’s experience these terms of service are not well enforced.

He cited a book by David Patrikarakos called “War in 140 Characters: How Social Media Is Reshaping Conflict in the Twenty-First Century.” The book argues that on social media, emotive images and narratives carry far greater weight than facts and rational debate.

“If we take this particular conflict, the images that Hamas has an interest in circulating are images of dead children,” said Ratson.

He pointed to a colleague of his, a professor at Oxford University, who often shares images of civilian suffering in Gaza accompanied by calls for Israel to be held accountable for war crimes.

“The professor’s interest in the human suffering in Gaza may have had honest authentic beginnings. But the more she signaled interest to the social media algorithms, the more Gaza suffering content was served to her by the algorithms,” he said.

“It is not far-fetched to assume that the more she signaled interest, the more the algorithm fed her with the same content until eventually she became part of a network that circulates those images herself,” he said. “She’s now an active voice of the Palestinian cause. She’s blindly anti-Israel. This is the type of behavior change that is facilitated by social media.”

read more:
comments
Never miss breaking news on Israel
Get notifications to stay updated
You're subscribed