The U.S. election and the rise of fake news

The introduction of new technology most certainly has an effect on the election cycle and the U.S. democracy at large. When the television was first invented and became a household appliance, people worried that it would reduce serious policy debates to nothing more than clips on a screen and promote telegenic politicians over those who are capable of running the country. Then the internet was invented and with it came the introduction of sourcing news from online, prompting concerns that it would be easier for people to create "bubbles" and surround themselves with only information they agree with.

Both of these technologies relied on some sort of third-party fact checking organization or person, a job that organizations welcomed and didn't take lightly. The shift to social media platforms like Facebook and Twitter, however, removes this middleman and creates an unfiltered forum for real and fake information to spread. Content can spread across screens, countries and continents in a matter of hours, and in many cases organizations with bad intentions can reach as many readers as reputable sources like the New York Times.

The issue of false news has only become an issue in recent years, but for those who are still uncertain, here's a couple facts to sway your mind. Estimates point the number of people who get their news on social media websites, like Facebook and Twitter, at 62 percent and collectively, the most popular fake news stories got more shares and likes on Facebook than actual news stories. To make matters worse, fake news stories are often believed to be true according to an Ipsos poll.

Predictions

During the 2016 U.S. election the public got an insight for the first time of how the media actually calculates their predictions for the election. There’s a lot of math and science that’s supposed to go into it, but essentially a small amount of people are asked who they will vote for in hopes that it’ll represent the choices of the whole country.

The way that many polling companies determine who is more likely to win is through random polling, where a worker simply calls a random number to see who they will vote for, detailed GOP operative Ned Ryun. They hope this way of gathering sample data is representative of the country as a whole and that all people called will be honest about their choice. However, this isn't always the case.

Claudia Deane, VP of Research at the Pew Research Centre, explained that some groups of people, such as those with a lower level of education, will be less likely to answer the phone. Less-educated white people generally supported Trump more, a sign that polling companies were likely to be less accurate from the get-go. People who were under the impression that their answer could lead back to them threw off the polls too. Many controversial things were done by the Trump campaign on the trail and people might not want to associate themselves with these acts, which could reflect back badly on themselves.

The other reason why the polls could have been so wrong is that they are meant to ask only people who are going to vote who they will vote for β€” a lot of the time, these people don't end up voting. The campaign was an odd one this year: both candidates had high unfavourable ratings and an overwhelming amount of voters were angry towards them.

There's a nonresponse bias built into this sort of polling method, explains the Pew Research Centre, which occurs when people don't respond to surveys despite equal opportunities for everybody. Groups like Trump's supporters of which the majority was less educated are harder for polling companies to reach, coupled with the anti-institutional and "fake media" feelings that were at the centre of his campaign. Quite possibly, the result of this could be a large amount of pro-Trump supporters not truthfully or at all telling pollers their voting plans.

James Lee from Susquehanna Polling & Research Inc. explained that his polling firm uses a combined method of live-interview and automated calls, and that Trump supporters were more responsive when sharing their intentions during the automated calls. There's no solid explanation for this, but it could be because voters are distrustful of the institutions Trump has been criticizing all along, or might not be comfortable talking about their political views with someone.

It is evident that the media lacks an accurate way of predicting polls just by looking at the predictions. On the morning of election day, FiveThirtyEight put Clinton's odds at a 71 percent chance of winning, but by the end of the day had flipped and calculated that Trump's odds of winning were 84 percent.

Fake News

 

Veles, a city with a population of 43,000, used to be known as a town that made porcelain bowls and plates that were sold throughout Yugoslavia, when it was still a country. Now the Macedonian town is known for its ability to churn out bogus stories designed to get the attention of and to trick Americans.

The scale of the operations in the city is large and by the end of the 2016 election, upwards of 140 websites could be traced back there. These websites have names like WorldPoliticus.com, TrumpVision365.com and USADailyPolitics.com, which all sound as though they might be real websites. But the news on the homepage was inaccurate and misleading, but somehow managed to end up in the feeds of millions of readers on the other side of the world. The city of Veles was full of masterminds who had managed to trick the public into believing all sorts of lies β€” and it was very clear that this town favoured Trump over the alternatives. 

In a CNN profile of several people from the town it became quite apparent that these people seem to be pretty ordinary, coming from backgrounds of all sorts and creating these fake websites to cash in on them. Google failed to block the websites from its ad network, allowing the creators to make thousands of dollars a month. The incentive for creating these websites with a target of Americans is simple: people in the U.S. earn up to four times more than non-U.S. people who click on ads. These people are turning to Facebook to make a living and the company has been doing nothing to stop that from happening.

The websites created in this city don't just pertain to the election; they cover websites containing anything from false health news to celebrity gossip, doing it simply because of the money that can be made. Yet somehow, there is no laws being violated and no remorse is felt for deceiving so many people.

Throughout the election both candidates repeatedly condemned Russia for attempting to influence it, but neither has mentioned the numerous Twitter bots that work on their behalf. The Atlantic uncovered two bots on the platform in support of both Trump and Clinton. The first one, @amrightnow, has over 40,000 followers and was created to tweet anti-Clinton conspiracy theories, while the second one, @loserDonldTrump, was made to retweet all mentions of Trump's official account containing the word "loser". These two accounts represent only a small fraction of the accounts on the platform that could have influenced the election.

One main issue of this whole fiasco is that the Federal Elections Commission doesn't recognize the existence of bots that are used to spread hate speech, harass voters and spread misleading and fake news. Bots are designed by people, but are created with the intent of being anonymous, which fails to hold them responsible in cases like this. It creates a need for regulations surrounding this type of technology in elections and though companies are getting smarter at tracking and taking down this sort of accounts, their creators are getting smarter too. Twitter and Facebook can already flag suspected accounts, such as those that tweet several times a minute and constantly tweet from the same IP address, but bot designers told The Atlantic that they can easily get around the detection methods.

Links posted during the election period from Nov. 1-11

University of Oxford researchers analyzed seven million tweets posted in the U.S. between November 1-11 and nearly one million of the tweets contained links to political content. Twitter users were exposed to more misinformation than reputable news, with the former numbering 256,725 links and the latter 259,342.

Facebook is home to fake news in the same way as other platforms, but over time it's evolved into an echo chamber for users, which poses its own issues. The whole idea of the platform is to connect you with people you already know, who most likely share the same political beliefs and ideas as you. The News Feed, where most content is seen, is constantly being filled with posts and news articles that conform to your interests β€” it is not meant as a place where ideas are supposed to be challenged.

In response to the rising concerns around social media and elections, Facebook Canada is launching a publicity campaign in advance of the next federal election. The initial phase includes the Cyber Hygiene Guide for politicians and political parties, which consists of training and guidance to safeguard against security threats and the launch of an email hotline. The social network said it's working to crack down on attempts to spread false information on the platform.

Social Media

The rise of social media over the last few years also certainly gave influence to what happened in the election. Both candidates had more online tools, such as Facebook, Twitter and Instagram but also Periscope and other upcoming networks to reach their audiences in ways not possible at this level in previous elections.

In August, Trump took to Twitter to thank a fan named Nicole Mincey, according to a report from The Independent. Several days after the interaction the account was suspended, because it was suspiciously similar to a series of other accounts linked to a Russian misinformation campaign on the platform. The incident gives insight into Trump's unfiltered tweeting and how it hasn't changed, despite his new role as President. Though it's not ordinarily someone's job to sniff out fake accounts, with that job belonging to Twitter, but as President of the United States, someone might've assumed that he has someone vetting his social media posts.

Social media could have been used to further the message of both candidates fairly and accurately, as it was in most cases. But, thanks to an Oxford University study, we found out that during the first and second debates up to one-third of the pro-Trump and one-fifth of the pro-Clinton tweets came from automated accounts. FiveThirtyEight published an investigation into the number of fake followers both candidates had on the platform, claiming 8 percent of Trump's and 7 percent of Clinton's followers are fake. The investigation was based on questionable characteristics of the follower's accounts β€” ones with few tweets or followers, or those that only joined the platform a few months ago.

Throughout his campaign Trump received a lot of attention for the poor quality of his racist, sexist and generally uncensored content, but that's exactly what his strategy throughout the whole election seemed to be. He used his Twitter account for unfiltered communication with his followers, numbered at many more than his competitors Clinton and Sanders. There's not a specific definition or box of what a fake account would fit into, but it's suspicious to see that a significant amount of his followers were new to Twitter, had less than one hundred tweets and primarily or only tweeted pro-Trump content. The Washington Free Beacon roughly estimated that as little as 2.2 million of Trump's then-7.51 million followers were fake and that between November 2015 and April 2016, 64 percent of his followers were inactive.

In Somalia, Somaliland has announced that it will restrict access to Facebook, Twitter, LinkedIn, Snapchat, Instagram, Google Plus and more in order to limit both hate speech and fake news in its presidential elections. Due to a lack of ability to control what is spread on the social networks, phone companies blocked the websites from November 13 until the votes were declared. Uganda did the same thing by blocking Facebook and Twitter during the time when its polls were open, but did not publicly announced it had done so.

Later on it was revealed that a Russian troll farm spent $100,000 on 3,000 Facebook ads, targeting U.S. citizens voting in the election. That hundred thousand dollars sounds like a lot, and that's because it is. For that much money, assuming the spend per ad is $33, tens of millions of people could be reached.

The narrative of Facebook and the election is imperative to a tale like this because, well, Facebook is the media now. It's introduced things like Instant Articles and hosted video in a way that turns it into an indirect news publisher, making the lesson that traditional news outlets like The New York Times and FiveThirtyEight learned one that it too must learn from. The concept that a group of people influenced the election in such a drastic way is still a foreign concept to most people; however, the media needs to learn from this series of events in such a way that it can prevent them from happening again.

πŸ“Έ: Wikipedia πŸ“Š: University of Oxford