Facebook to remove Israeli political bots

Facebook and the Israeli elections Photo design Tal Bogdansky

In efforts to reduce its influence in the upcoming elections, Facebook is also demanding information on how much each candidate is investing in each post and the audience they aim at.

Elections in Israel are just around the corner, and Facebook is redoubling its efforts to prevent the use of its social network to influence them. Sources inform "Globes" that Facebook's campaign to thoroughly clean up accounts of the followers of politicians will reach a peak in the coming weeks. Fraudulent and automatic accounts (bots) will be removed from the network and the pages of the parties' candidates. The network is also offering tools to those dealing with media for reporting networks of fraudulent users.

This campaign comes on top of the tools that Facebook plans to launch next month in Israel involving political awareness. These tools will include identification, a link to published pages, and preventing the purchase of such notices by overseas parties. At the same time, Facebook will hold a closed event next week for political campaign managers in which rules for election campaigns will be presented. Among other things, Facebook is demanding information about how much each candidate is investing in each post and at what audience it is aimed.

Facebook has taken similar measures in the past two years in election campaigns throughout the world, following severe public criticism concerning the 2016 US presidential campaign. At the same time, marketing sector sources were surprised when we asked them about Facebook's "bots campaign," saying that they had not been informed of it. It is believed that there are hundreds of thousands of forged accounts on the social networks in Israel, at least some of which are promoting political parties and agendas.

Sources in the sector say that one of the politicians likely to suffer significant damage from the measure is Prime Minister Benjamin Netanyahu, who emphasizes digital means of influencing public opinion. Another potential loser from the measure is Labor Party chairperson Avi Gabbay. It was revealed last June that Gabbay's supporters had operated a Facebook page aimed at attacking Yesh Atid chairperson MK Yair Lapid. "Netanyahu and Gabbay have the most money to fund the election campaign: Netanyahu's party has 30 MKs and Gabbay's has 24. They are the only ones who can spend millions on these accounts, because it's ineffective if done with small amounts of money. None of the others are doing it, at least not on a level that can change the discourse," a source said.

Mayan Sarig, Facebook's spokesperson in Israel, told "Globes," "We're working hard to keep our platform clean all of the time, including the removal of phony profiles. This is an ongoing measure that takes place all year, and certainly now, since elections were announced in Israel."

Objective of the bots: To create an illusion that something has gone viral

The struggle to eliminate phony accounts from Facebook is part of the war against fake news, which reaches a peak during election campaigns. In the past, the social network assumed that the problem was fairly negligible, despite evidence that it was having an effect as early as Russia's annexation of the Crimean peninsula in 2014, and later in the Brexit campaign in early 2016. Even after the election of Donald Trump as US president in November 2016, Facebook founder Mark Zuckerbeg claimed that the idea that fake news had affected the results was a "pretty crazy idea."

The accumulation of evidence about Russian involvement in the online campaign to elect Trump, however, began to cause real damage to Facebook's image, leading to a change in the company's attitude. Facebook took steps to thwart fake news in a series of elections campaigns around the world in the past two years, including the French presidential elections in 2017 and the US Congressional elections last year, which were believed to be relatively clean. Like its rival, Twitter, Facebook is careful to report the number of accounts that it removes in the framework of these efforts.

The fictitious users are replacing the talkback armies used by political parties in the past during election campaigns - the keyboard armies that linked the party's messages page to responses on the news websites. The bots scan the Internet, spot conversations on the topics relevant to their operators, and respond automatically according to the message that they are programmed to deliver. In many cases, they quote posts written by real accounts in order to avoid being identified. They also work in tandem with avatars - political party workers impersonating average people - and accounts that combine the two methods.

The problem is not confined to messages distributed via phony accounts. For example, take a surfer exposed to content posted by a given candidate. He sees that the content received hundreds of likes or other emoticons (hearts, smiles, laughs, and so forth), but does not know that phony accounts are responsible for half of them. The impression of an event and a viral and interesting message is created, and it is more likely to reach a larger audience. The opposite is also true: if a certain candidate is flooded with negative responses and angry emoticons, the surfers will not know whether real people or fraudulent accounts are behind them. Marketing and advertising firms use these tactics for the politicians they promote every year, but the situation gets worse during election campaigns.

The subject became important in the US after it was learned that an enormous farm of bots and computers were used in the elections to create fake content and news that never happened, and to publicize and promote it. The claims against Trump concerning Russian financing were based on this very issue: the belief that his presidential campaign included millions of phony accounts on the social media and the Youtube video website. The law authorities are currently investigating who promoted the fake news campaign, to what extent the Kremlin was involved and financed it, and what the president and his associates knew about it.

The result was severe damage to Facebook's image, which for years has boasted of being a "platform of real people." The trust of users and advertisers in Facebook was affected. The attitude towards it as a pioneer of democracy and free thought, which was prevalent during the Arab spring, changed completely, and various countries began to consider restrictions on it, mainly involving its use of private information and how it uses targeted announcements. The bots and avatars are also directly costing Facebook revenue. In contrast to the purchase of advertising to influence surfers, Facebook earns nothing from the activity conducted through fake accounts.

Detection of phony accounts is largely based on algorithms using artificial intelligence to assess irregular behavioral patterns. A user with few friends and posts, all of whom focus on a specific subject, is likely to arouse suspicion. The same is true of a user posting a like for every post by a specific politician and doing nothing else on the Internet. Other phony accounts use more sophisticated methods, but the ability to distinguish them from real users is constantly improving.

These technologies obviously require the investment of large resources and amounts of money by Facebook, but the company concluded that it had no choice other than to do so, and hired many thousands of human content checkers. This effort is not confined to the political arena. Surfers were recently removed from Facebook groups that had not joined the groups on their own initiative. Some groups had many thousands of such users. The company previously took action against brand pages that paid for likes.

Goal of the websites: To filter and preserve revenue

Facebook is not the only digital unit having to fight to maintain the integrity of its platform during election campaigns, although it is the most powerful platform in terms of its ability to shape awareness. International and local concerns face similar challenges, and are striving regulate political campaigns effectively, while minimizing the resultant loss of revenue as much as possible.

Two weeks ago, "Globes" revealed that Google had issued tough detailed guidelines restricting political advertising during election campaigns. Among other things, targeting on the basis of demographic and behavioral profiling (the focus of the Cambridge Analytica scandal, in which Facebook was embroiled) and the use of e-mail and retargeting (advertising based on the user's previous behavior) were banned.

Israeli content recommendations companies Taboola and Outbrain, which were used to spread defamatory information about Knesset candidates, tightened up their rules and notified the media companies that they would allow only positive content, not negative campaigns. The need for full disclosure was also stressed.

Content websites in Israel are also being more careful about filtering, but only up to a certain point. No one even dreams of foregoing revenue from the elections. Starting early next months, the political parties will put more money into campaigns and make them more strident. As a result, most of the websites have issued a special price list for the elections, with substantially higher rates than those charged at ordinary times. One of the concerns that has instituted such a rate is the Artimedia programmatic advertising network.

The Likud said in response, "The Likud and Prime Minister Benjamin Netanyahu do not operate even one bot."

Published by Globes, Israel business news - en.globes.co.il - on February 21, 2019

© Copyright of Globes Publisher Itonut (1983) Ltd. 2019

View comments in rows
Update by email about comments talkback
Facebook and the Israeli elections Photo design Tal Bogdansky
Facebook and the Israeli elections Photo design Tal Bogdansky
Twitter Facebook Linkedin RSS Newsletters גלובס Israel Business Conference 2018