How a Report on Israel's 'Right-wing Bot Network' Taught Me a Lesson on the Bot World

This story was originally titled- 'An Israeli investigator revealed a right-wing bot network, and it may be the tip of the iceberg.' But in the meantime an alternative explanation emerged, suggesting that the network may be international. What's the answer?

Send in e-mailSend in e-mail
Send in e-mailSend in e-mail
A fake Twitter profile found by investigator Noam Rotem.
A fake Twitter profile found by investigator Noam Rotem.Credit: Noam Rotem
Oded Yaron
Oded Yaron

UPDATE: Haaretz reported Monday that security investigator and activist Noam Rotem exposed on Twitter a bot network which he estimates is being activated by political sources in Israel and mostly tweets information in favor of the right wing in Israel.

The person behind a Twitter handle named jewish_warriors who also followed the network believes that it is actually part of an international bot network that operates in several languages including Korean, Arabic, English and Spanish.

JW started following the network after he recognized the similarity between different tweets written by several Twitter handles he recognized as bots during the World Cup games. He believes that that the network is being operated by sources in the United Arab Emirates.

I have since tried to corroborate the information that JW has compared with the information Noam Rotem collected to determine which version is more accurate; I'm still not entirely sure.

While it seems that the two conducted honest and serious research, and it doesn't seem that either of them tried to tip the results in any direction, they reached different conclusions.  In fact, Rotem himself referred to other bot networks but his assessment was that they are operating simultaneously—he didn't link them to the specific network he was researching. JW, on the other hand, found evidence to link between the networks.

This is the original report: 

The 2016 American elections shed a spotlight on various efforts to influence the political system using bots – software applications that can, among other things, mimic the responses of real people. In recent months, Israeli users have started to get the feeling that there’s been an increase in activity by false profiles on social networks – particularly on Twitter, but on Facebook as well – and on Sunday proof was offered of what looks like a bot network that is disseminating political tweets.

To really understand Israel - subscribe to Haaretz

Security investigator and activist Noam Rotem explained in an online post that he had started looking into the issue following complaints by users that they seemed to be surrounded by bots. “A few weeks ago I asked friends to send me a list of [profiles] they thought looked suspicious and I got hundreds of them,” he said. “Of those hundreds, I succeeded in confirming through various means that these were real people, not bots. That’s not always easy, because there are some people who aren’t, shall we say, the sharpest knives in the drawer, but I reached a reasonable level of certainty that they weren’t bots.”

A fake Twitter profile found by investigator Noam Rotem.Credit: Noam Rotem

But Rotem also discovered a few different networks, at least some of which were connected or dealt with similar topics, and succeeded in closely following one of them, which he said is made up of “dozens of bots that tweet primarily about politics from the right side of the political map.” The accounts he identified monitor each other, and when one of the accounts tweets, the others follow.

 “Most of the accounts pretend to be young women, some of them soldiers, which between one political tweet and another publish pictures of our country’s landscapes and well-crafted Shabbat Shalom messages,” according to Rotem. As in the case of many other fake profiles trying to attract likes, he said, “each and every one of them shows a picture of a model or an internet star, some of whom prefer to advertise themselves wearing very little, if any, clothing.” Rotem added that there’s even a certain dissonance between the revealing photos and the central place that tradition and religion occupy in the accounts’ tweets.

>> How a 29-year-old 'Russian agent' used a pro-guns agenda to infiltrate U.S. conservatives ■ The countless Israeli connections to Mueller's probe of Trump and Russia | Analysis

Rotem’s latest discovery comes after developer Ran Bar-Zik identified a bot network a few months ago, and after Facebook admitted (after many twists and turns) that right-wing activist “Adam Gold,” whose page was followed by many people, including journalists, was actually a fake profile. The bots Rotem found in his research weren’t enormously popular nor did they try to achieve prominence like Gold (which in the end proved to be the undoing of those operating the page) but were part of a network that was quite amateurish and easy to map, in his words.

This network tried to help spread the opinions of its operators and to “intensify the message,” as Rotem put it. “When you see a tweet with zero retweets and zero likes, you’ll relate to it differently than a tweet that already has a hundred retweets and 600 likes,” he explained. Bots of this type are also used to attack users or politicians from the “wrong side” of the political map or to divert a discussion in the preferred direction.

Thus, for example, Rotem said the accounts were used to spread reports of a race-based attack and to recycle the claim that the left was responsible for financing various protests, like that of the Druze against the nation-state law. As was reported, over the weekend there was a forged WhatsApp exchange disseminated that made it look as if the Saturday night demonstration was the result of deals made between Druze leaders and Labor Party activists.

Rotem used software to identify similar features among the profiles and found, among other things, that the network’s operator used an online campaign management tool called Social Report. This platform can automatically duplicate and distribute certain tweets via multiple accounts. According to the company price list, the service costs between $50 to $100 per month, depending on the number of accounts. However, Rotem says that the real cost of the operation could be thousands of dollars, taking into account the necessary work hours and other basic expenses.

The use of bots and fake profiles for political purposes is not new, with the first reports of this emerging at the beginning of the decade. But the recent elections in the United States, in which there was allegedly extensive Russian influence on the campaign along with many other questionable online activities, put the issue at center stage.

The fact that he could identify the source with such ease points to a pretty unsophisticated operator, Rotem said. The low number of accounts of in the network could mean that it is still nascent or that it’s some kind of trial.

Dr. Anat Ben David, head of the Media and Information Laboratory at the Open University, told Haaretz that while the network Rotem uncovered looks quite amateurish, more serious operators would not limit themselves to such a network, but would use it in combination with the many other tools available in the political arsenal.

She added that despite the feeling that this phenomenon is spreading, it’s hard to estimate its real scope or confirm it with solid data. One thing we can learn from these revelations, she says, is that concern should not be focused solely on possible foreign influence, since “political elements within the Israeli system are using similar methods to conduct problematic manipulations of Israel’s domestic discourse.”

Click the alert icon to follow topics: