The number of fake users on social media is rampant – and they have had more and more of an impact in recent years.

If Twitter’s congressional testimony is to be believed, about 5 percent of their active users aren’t real people. However, some independent studies, such as this one from Indiana University and USC, have pinned the number as high as 15 percent for their platform. That’s about as high as Facebook says their bot percentage was in 2017, in the wake of the US presidential election. They now claim that around 4 percent of their users are fake accounts.

Unfortunately, getting an accurate idea of the number of active fake users on such a scale is inherently a difficult process, and it’s more than possible that official figures are underestimations. In fact, Facebook even says as much in their recent SEC filing, where they state that

“It is possible that the actual number of duplicate and false accounts may vary significantly from our estimates.” -- Facebook Inc, April 2019

There’s plenty of reasons for fake users on the internet – whether they happen to be bots (the vast majority) or simply people pretending to be someone they’re not. Some of them are there to hawk a product of some sort, a lot like getting robocalls on your cell phone. Some are there to scam you out of your hard-earned money, by pretending to be your relative lost in another country or posing as royalty who will definitely wire you millions as soon as you give them your credit card number.

Even non-traditional social media, such as dating apps, have become inundated with fake users hoping to tease a buck out of their unsuspecting target – or in a recent case, the locations and images of Israeli soldiers. Still, most would agree that perhaps the most relevant purpose of fake users is manipulation of opinion, especially in relation to politics.

Social media can be an important vector for scams to proliferate

During the US 2016 elections, for example, political bot activity accounted for millions of posts on social media. At the peak of activity, they contributed to over a quarter of posts on Twitter, according to this study. Fake users have attempted to influence discussion around Brexit, the conflict in Ukraine, and the recent protests in Hong Kong. And as can be seen in the last example, governments have no qualms about taking advantage of these strategies as well.

The Chinese government, for one, has “long been suspected of hiring as many as 2 million people” to promote their narratives, and it’s been estimated that they “fabricate and post about 448 million social media comments a year,” by a Harvard study. Such an undertaking would require the services of a veritable army of bots and paid “trolls” – something that the Kremlin also knows quite well, having employed a shady organization known as the “Internet Research Agency” to do its bidding. Unfortunately, many users have been shown to be vulnerable to this type of manipulation, and end up reposting fabricated material, inadvertently amplifying the “fake news” even more.

People fall for fake users and their fake news at an alarming rate today. One reason for this may be that they associate some social platform metrics (like friend count or connections) with trustworthiness. Social media users tend to respond more to fake users with large friend counts, a study showed, and it is simple and inexpensive to buy or generate fake accounts to do so. Many users often wrongly assume that there is a genuine person behind each account as well, and that their “platform’s privacy settings will protect them from fake profiles”.

Perhaps most concerning of all, even the simple act of being exposed to fake news can make you more likely to believe in its accuracy over a period of time, even if it is marked as contested by fact checkers, or presents a view that conflicts with your political slant.

Online "trolls" can be directed to influence politics by governmental bodies

To their credit, social media platforms and companies aren’t just sitting back and doing nothing. Facebook has been taking down millions of pieces of spam and fake accounts every day, according to their own report, most of the time within minutes after post or registration. Many fake accounts are prevented from registering in the first place.  Twitter also proactively challenges millions of accounts per week on its platform, as does the professional social network LinkedIn.

Techniques used include relying on user reporting, and more commonly these days, the application of advanced artificial intelligence and machine learning methods to automatically detect fake users. As time passes, we can expect to see these methods improving. For now, though, we will have to put up with what gets through the filter, as this New York Times journalist found out. After all, even if only 4% of Facebook’s active users are fake, as they claim, that still amounts to 98 million accounts when compared to their 2.45 billion monthly active user count.

The cure for the epidemic of fake users, which is sure to only become more of an issue in the coming years, isn’t clear-cut and dry. Companies, such as Moonlight, are advancing ideas like more robust profile information sharing controls and blockchain-based verification as ways to help users have more confidence in the security and validity of their online interactions.

However, technological improvements to current methods (such as the aforementioned AI and machine learning), and development of new fake user detection methods or combinations thereof are only part of the solution. Users themselves must also be encouraged to take more time to carefully evaluate potential connections or news articles and learn how to identify them by their traits. Resources like Mozilla and Harvard have some great, simple guidelines on these matters.

Newspaper publishers Joseph Pulitzer and William Randolph Hearst were well known for pioneering sensational stories 

It’s important to note that fake news is not a new thing, despite its new moniker. From the rise of “Yellow Journalism” in the 19th century, to back when caveman Ogg spread unflattering rumors about caveman Ugg around the campfire, half-truths and untruths have percolated through channels visual, verbal, and otherwise. And while technology might have exacerbated the problem today, it may also have the potential to help undo it.

As a company in this space, we realize our responsibility towards the community and take it seriously. Going forward, we hope to engage in more conversation about subjects like these—and hopefully, make a difference through our work.