Avigdor Liberman, the Yisrael Beytenu party head and former defense minister, is a Russian spy. How do we know this? Because the former Mossad head Tamir Pardo said so, in an address to Harvard University’s Belfer Center in November, as reported on the center’s internet site. Except that Pardo, who did indeed speak to the Belfer Center, said nothing of the kind, and the center never claimed he did.
Yair Netanyahu, the prime minister’s older son, visited the UAE at the invitation of German manufacturing giant Siemens in October, met with the crown prince of Dubai, Sheikh Hamdan bin Mohammed Al Maktoum, and discussed the possibility of investing Netanyahu family assets in the Gulf states. How do we know this? Because the CFO of Siemens Middle East was quoted on the European Coatings website discussing Yair Netanyahu’s visit. Except that the prime minister’s son did not in fact discuss any such investment plans, did not visit Dubai, and was not the subject of any such statement by Siemens on any such website.
These are just two recent examples of fake news stories — things that didn’t happen and weren’t said — that were “reported” on cleverly faked websites.
They were then tweeted out from fake accounts to Israeli journalists in the hope that the reporters, duped, would launder the false stories by publishing them as genuine news items on genuine, widely read news sites.
Sounds worrying? Here are three more recent examples of false stories masquerading as genuine news being disseminated in this and similar ways:
A report that Hatnua leader Tzipi Livni complained in a recent speech at the Herzliya Interdisciplinary Center that the Mossad badly mishandled matters relating to the Khashoggi affair, with the consequence that a Saudi official who was at the heart of efforts to warm ties between Jerusalem and Riyadh was fired. (Livni issued no such criticism.)
A report that Israel’s ambassador in Sweden was engaging in mediation efforts with Yemen’s Houthis. (He was not.)
And, the Netanyahus again, reports claiming that Sara Netanyahu wears a hidden cross and that son Yair has converted to Christianity. (Needless to say: false information.)
All of these stories struck one or more of the people to whom they were sent as improbable, thus beginning the process by which they were exposed as fake and traced back to the accounts that first tweeted them. None of them seems to have made it into mainstream media. All of them had begun to gain traction, however, including being retweeted by some of their recipients to whole new audiences, before they were stopped in their tracks.
All were insinuated into the Israeli social media discourse from overseas. From where exactly? Only the social-media platforms themselves can say, since they have the IPs (computer addresses) of the fakers. All were designed to advance certain interests, including undermining this or that public figure.
Sophisticated efforts to harm Israel by disseminating fake news over social media have intensified since Israel moved into election mode, but they have been going on for years. A foundational case in point: In August 2015, Paris Mayor Anne Hidalgo hosted a one-day “Tel Aviv on the Seine” festival, in partnership with her Tel Aviv mayoral counterpart Ron Huldai. The event went ahead as scheduled, but only after then prime minister Manuel Valls backed it, and backed Mayor Hidalgo, in the face of what appeared to be a major, grass-roots outcry.
The “outcry,” in fact, was a sophisticated campaign of “astroturfing” — a purported viral surge in social media protest that was actually an orchestrated case of Twitter manipulation. This was revealed by a Belgian social-media researcher named Nicolas Vanderbiest, who established two days before the event that Hidalgo, several senior politicians, and a small group of leading French journalists were being bombarded with tens of thousands of anti-Israel tweets demanding that the festival be canceled. These were being automatically generated by clever but simple software that sent out tweets from, among others, the hijacked accounts of thousands of vacationing French citizens.
Convinced by this wave of tweeted protest that the event might prompt a dangerous surge in violence, involving pro-Palestinian activists and their supporters, the worried journalists penned front-page articles to that effect in the major French dailies, questioning Hidalgo’s decision to allow Tel Aviv its sunny day in the heart of Paris. The next day, major French TV stations picked up the theme, debating Hidalgo’s decision on news broadcasts and in talk shows.
On the day, actual protests against “Tel Aviv on the Seine” were so insubstantial that there was grumbling about the heavy-handed, inconvenient 500-strong police presence, which had been called out to handle what turned out to be no trouble. “The police presence, larger in number than the holiday-makers, gave an absurd and oppressive tinge to the event,” according to one report.
The “Tel Aviv on the Seine” astroturfing crisis, which diplomats say has affected cultural relations between the two cities to this day, was so potentially dramatic, and its faked origin so revelatory, that it prompted Yuval Rotem, the director-general of the Ministry of Foreign Affairs, to take the radical and fairly prescient step that summer of setting up an R&D department for “Algorithmic Diplomacy” in the ministry, headed by diplomat Elad Ratson, tasked with seeking to highlight and help thwart this kind of algorithmic manipulation of public opinion against Israel.
How the fakers do it
How was “Liberman’s a spy” launched into the social media maelstrom?
With devilish simplicity. As detailed by Haaretz in November, the fakers duplicated the website of Harvard’s Belfer Center, at belfercenter.org, purchased the deceptively similar belfercenter.net domain, set up a spoof (hoaxed replica) of the original site, and published the fake article where they falsely quoted Pardo as revealing, “In 2011, when Lieberman served as Foreign Minister, we obtained pieces of evidence showing his secret ties with the Federal Security Service of Russia…”
The URL with the fake story was then tweeted out from a Twitter account in the name of “human rights activist” and “martial arts” enthusiast “Bina Melamed” — an established account that appeared credible — to various Israeli journalists. (The “Bina Melamed” Twitter page has since been closed.)
Similarly, the “Yair Netanyahu in Dubai” claptrap was first “reported” on a spoofed version of European Coatings’ website. The fakers replicated european-coatings.com on a domain they purchased from Colombia — european-coatings.co — and published their false account there of Yair Netanyahu’s Dubai dealings, under the headline “UAE; Middle East Investment Destination for the Israeli Prime Minister’s Son.”
The URL with that fake story was then tweeted out by one Mazal Shapiro, a “freelance journalist” and “academic researcher” who engaged Israeli reporters’ interest by asking them whether they had seen the story. (As of this writing, Shapiro’s Twitter page is still up, with her denial that she knowingly shared fake news.)
The Livni-Khashoggi lie was disseminated to Russian-speaking journalists. The Netanyahu Christianity fakery was tweeted out in Google-translated Hebrew, via seemingly credible Twitter accounts in the ultra-Orthodox community. The Israeli Ambassador to Sweden meets the Huthis fake was spread via a spoofed Russian-language Israeli news site — the genuine cursorinfo.co.il replicated at cursorinfo.co — and then, again, disseminated to Israeli journalists via a seemingly credible Twitter account.
All these fakes were stopped in their tracks thanks, first, to journalists who smelled rats, and then to a coalition of researchers, high-tech companies and the Ministry of Foreign Affairs’ Algorithmic Diplomacy team, who traced the origins of the false stories, pinpointed the bot and sock-puppet networks which gave them the appearance of genuine news, and provided Twitter with the necessary proof to thwart the abuse.
The Foreign Ministry’s department, explains Elad Ratson, works with various computer science experts, data scientists, tech companies and university researchers, in Israel and in Europe. These experts look for anomalies in social media, and share information on what they find. It was a researcher named Ran Bar-Zik, for instance, says Ratson, who identified the spoofing of the Harvard University Belfer Center site and the Bina Melamed Twitter account which seeded the fake Pardo-Liberman-spy story among Israeli journalists. It was a company named communit360.com which pinpointed a key 80-Twitter account bot network that helped hype the story. And it was software developed by the ministry that tracked down the Israeli journalists who were targeted by that network; the ministry then liaised with Twitter through the process to provide the proofs required for the permanent suspension of dozens of accounts associated with the incident.
Unsurprisingly, the social media platforms don’t like to be seen as being abused by illicit misinformation campaigns, particularly when such campaigns are state-orchestrated or state-linked. But it is only the platforms themselves, of course, that can take down the fakers’ accounts. A little hint of the discomfort may have been discernible only last week, when Ratson tweeted that Twitter had suspended “yet another batch of 61 accounts with a total of 28,041 followers, all linked to foreign #FakeNews manipulation campaigns aimed at Israeli public.” A Twitter spokesperson offered a somewhat different narrative, retorting sniffily: “We have carefully reviewed these accounts and our analysis does not indicate any malicious intent or State-sponsored activity.” Nonetheless, the accounts remain suspended.
Tackling the fakers, plainly, is a complex and expensive process. In the case of the manufactured anti-Israel outcry over “Tel Aviv on the Seine”, the hyped false hysteria was traced to a total of 39,688 tweets, generated within three days and attributed to 10,428 Twitter accounts –of which only 2,941 could be traced to genuine French users. The rest were bots and sock-puppets, which created a purported avalanche of protest that managed to set the political agenda in Paris in the run-up to the festival, and came close to achieving the goal — the cancellation of a pro-Israel event.
More recently, in the case of New Zealand pop singer Lorde’s decision in December 2017 to cancel a concert in Israel under ostensible huge and vast BDS pressure, 96% of all Twitter traffic directed at her, in the successful effort to pressure her into cancelling, was ultimately traced to 172 Twitter accounts, I was told. “If she had known that, perhaps she would not have been deterred from performing,” mused the expert who gave me those figures.
If Lorde had known that 96% of all Twitter traffic directed at her came from 172 Twitter accounts. ‘perhaps she would not have been deterred from performing’
The process of “reverse-engineering” a superfast-spreading “astroturfed” campaign — that is, working back to identify both the original fake tweets, and their exponential circles of automated-retweeting and ongoing engagement — requires both expertise and access to information that Twitter does make available, but at a cost. To this day, not all of the fake Twitter accounts that were identified as disseminating the fake stories I’ve cited have been closed.
Sometimes, they shape-shift to try to stave off detection and closure. The “Bina Melamed” Twitter account, once exposed, for instance, transformed into “Leakers Without Borders,” the Foreign Ministry’s Ratson noted, and managed to fend off Twitter suspension by a few more days. (A similar process, Ratson added, played out with a Twitter account claiming to belong to ex-Shas MK Haim Amsalem which, when exposed as fake, changed its name last month and has fallen off the radar, for now.)
So much for Twitter. When it comes to Facebook, the task of spotting and blocking fake news is still more complicated, with potent implications for the current Israeli election campaign.
Facebook and other social media platforms were savaged by the US Senate over the way they allowed themselves to be strategically abused by the Russian state-supported Internet Research Agency (IRA) to meddle in the 2016 US presidential elections — a campaign of nothing less than “information warfare against the United States of America,” as a recent report for the Senate noted.
That December 2018 report established that the IRA carried out “a sweeping and sustained social influence operation consisting of various coordinated disinformation tactics aimed directly at US citizens, designed to exert political influence and exacerbate social divisions in US culture.” The IRA disinformation campaign, with a budget of over $25 million, reached 126 million people on Facebook, and at least 20 million users on Instagram — which is owned by Facebook.
Facebook’s owners and top staff, from Mark Zuckerberg on down, have been very publicly contrite, and have promised to fix the platform’s vulnerability to outside players who abused it to meddle with US democracy: “Foreign actors conducted a coordinated and sustained effort to attack our democracy… [and] abused our service,” it acknowledged last February. “We know we have more to do to prevent against future attacks.”
As of this writing, Facebook, outrageously, is quite simply allowing itself to host political ads relating to Israel’s elections, and rake in the money from advertisers, without those advertisers being required to identify themselves
But December’s report to the Senate made explicit that Facebook continues to stall researchers who are investigating and seeking to prevent further iterations of that strategic, Russian-directed abuse. Facebook “provides researchers with very restricted access to publicly valuable Facebook platform data through its Application Programming Interface (API),” the report complained, and “it currently provides none on Instagram.”
While the Israel-related fake news stories detailed above were planted on fake sites and pushed to journalists and other potential opinion-shapers via Twitter, researchers have established that another highly cost-effective means of spreading misinformation is via Facebook ads. As of this writing, Facebook, outrageously, is quite simply allowing itself to host political ads relating to Israel’s elections, and rake in the money from advertisers, without those advertisers being required to identify themselves.
Likud last month inexplicably rejected a plea from the Central Elections Committee to apply basic transparency standards to online campaigning
Last week, with some fanfare, Facebook announced that it would be introducing tools intended to prevent Russian-style meddling and manipulation in April’s elections, and other elections worldwide, by barring anonymous political ads. In other words, when Facebook users are urged in ads on the platform to support or oppose candidates or parties, they’ll be able to identify who is doing the urging. This is especially vital in the April elections since Netanyahu’s ruling Likud party last month inexplicably rejected a plea from the Central Elections Committee to apply basic transparency standards to online campaigning — inexplicably, that is, as Israel Democracy Institute expert Tehilla Shwartz Altshuler noted, unless Likud is planning to make use of some of the dubious Russian-style methods that gained prominence in the 2016 US elections. (Likud was the only party in the outgoing Knesset to refuse the Central Elections Committee request.)
But if Facebook’s announcement appeared to indicate that the social media platform was learning its lessons, and was truly determined to stamp out disinformation and meddling, the small print told a different story. The new rules barring anonymous ads in Israel are only set to come into force in March — toward the end of the campaign. Asked by The Times of Israel if it was even prepared to specify when exactly in March, Facebook maintained its routine haughty silence.
In the Israeli context overall, Facebook is quite simply a closed box. Its officials in Israel have insistently refused repeated requests by The Times of Israel for interviews about company policies and the measures it takes to defend against political meddling, disinformation and other abuse — whether engineered from inside or outside Israel.
And it refuses to disclose which pages it takes down at the request of the National Cyber Directorate, the agency based in the Prime Minister’s Office responsible for all aspects of cyber defense in the civilian sphere. Last October, ahead of the municipal elections nationwide, for instance, Facebook took down thousands of fake profile accounts created to spread false information about Israeli political candidates at the request of the directorate, but neither the directorate nor Facebook has been prepared to provide any specifics about what comes down, what stays up, and according to which criteria. (Facebook’s head of policy and communications in Israel is Jordana Cutler, a former top aide to Israel’s US Ambassador Ron Dermer, and a member of the Likud campaign team in the 2009 elections.)
Won’t be fooled?
All of us, this writer emphatically included, would like to believe that we’re savvy and smart consumers of news and information. Doubtless many of you said to yourselves, when reading my descriptions of what you now know to be fake stories about Liberman the spy and Yair Netanyahu the Christian, that you’d never for a moment have believed stories like that if they popped up on your Twitter or Facebook feeds. Indeed, you might not.
Are you completely certain that, if you read somewhere seemingly credible that the former head of Mossad had called him a spy, you might not have second thoughts about Liberman?
But when the misinformation is both relentless and carefully targeted, it does have a proven effect.
In a campaign designed to encourage Britons to vote for Brexit — leaving the European Union — “Russian and Iranian internet trolls sent more than 10 million Tweets… including a day-long blitz on the day of the Brexit vote,” Britain’s Telegraph reported in October, citing Twitter data. “Russian ‘troll factory’ the Internet Research Agency and a separate group of Iranian hackers used networks of Twitter accounts to spread divisive information,” the report said, adding, in echoes of what’s been happening in the Israel sphere, that “some of the posts came from accounts masquerading as news organisations and journalists.” Did the “troll factory” campaign have any effect? In a May, 2018 working paper, the US National Bureau of Economic Research estimated that Twitter bots “added 1.76 percentage points to the pro-‘leave’ vote share,” Bloomberg reported — an immensely significant figure, given that Brexit was approved by a narrow 51.9-48.1 percent.
The same National Bureau of Economic Research paper estimated that Twitter bots “may explain 3.23 percentage points of the actual vote for Trump in the US presidential race,” the Bloomberg report said.
But for a bitterly tangible example of plainly ridiculous fake news nonetheless gaining credibility and prompting actual, violent consequences, look no further than Pizzagate — the conspiracy theory spread via social media and fake news websites that Hillary Clinton ran a child sex trafficking ring from the basement of a Washington, DC, pizzeria.
An actual consumer of that insane misinformation went out and shot up the pizzeria.
Are you completely certain that, if you read somewhere seemingly credible that the former head of Mossad had called him a spy, you might not have second thoughts about Liberman? Somewhere deep down? Enough to plant a seed of unease? Enough to, just maybe, make you think twice about voting for him if you’d been hitherto contemplating the idea?
We have been warned
For a couple of days last month, Hebrew media focused heavily on a warning issued by the head of the Shin Bet domestic intelligence agency that “a foreign country” intends to intervene in Israel’s elections via online meddling. Nadav Argaman actually specified which country he was referring to, and said he was 100% certain of the looming effort, based on concrete information. Russia, unbidden, hurried to assert that it had no such intention — indeed, that it never has and never will carry out any such intervention in anyone’s elections.
Israel’s military censor continues to bar journalists from specifying whether Argaman was indeed referring to Russia. The merits of that censorship are debatable. What is not debatable is that Israel’s cyber directorate, whose purpose is to defend Israel’s civilians from online attack, and other relevant bodies, need to be complementing the Foreign Ministry’s efforts to ensure that the intervention fails. Central to providing that assurance would be investing whatever resources are necessary to enable the speedy identification and thwarting of fake stories designed to skew our preferences in the run-up to the vote, and the reiteration of the demand that platforms such as Facebook institute the thoroughly democratic requirement that the funders of all political advertising be identified.
Instead, the spokesman of the Central Elections Committee acknowledged to The Times of Israel last month that it was trying to devise a plan to guard against meddling, but doesn’t actually have the tools to do the job. For his part, the judge who heads the committee has been reduced to issuing pleas to safeguard the democratic process.
After Likud last month rejected his request to agree to apply at least basic transparency standards to online campaigning, Supreme Court Judge Hanan Melcer was forced this week to appeal to Facebook to bring forward its ban on anonymous political advertisements and expedite the introduction of other tools and restrictions aimed at preventing foreign interference in the elections — the supervisor of Israel’s elections making plaintive submissions to a social media behemoth that has acknowledged facilitating the abuse of elections elsewhere, and apparently sees no urgency in preventing the self-same abuse being perpetrated here. True to stone-walling form, Facebook merely promised to give him an answer in the near future.
What else is out there, masquerading as factual reporting, that hasn’t yet been flagged as fake?
I began this article with specific examples of fake news pieces that were spotted and discredited before they entered the mainstream and started to mess with people’s heads. What worries me at the end of it is the question of what else is out there, masquerading as factual reporting, that hasn’t yet been flagged as fake — that is, in other words, messing with our heads right now.
And what should trouble all of us Israelis, as we prepare to exercise our voting rights, is that Israel remains vulnerable to the ongoing online abuse of our democratic process. As the examples we know about underline, it’s already happening. This despite the fact that we know Russia planted false stories and disinformation that messed with voters’ heads in the US elections, the French presidential elections, and the UK Brexit vote. And despite the explicit warning from our domestic intelligence chief that the very same thing has been plotted by a foreign country for our electorate too.