Facebook Bots

[Berniece Leonhardt]

Thestaggering figure of more than three billion fake accounts blocked by Facebook over a six-month period highlights the challenges faced by social networks in curbing automated accounts, or bots, and other nefarious efforts to manipulate the platforms.

Facebook said its artificial intelligence detects most of these efforts and disables the accounts before they can post on the platform. Still, it acknowledges that around five percent of the more than two billion active Facebook accounts are probably fake.

Bots played a disproportionate role in spreading misinformation on social media ahead of the 2016 US election, according to researchers. Malicious actors have been using these kinds of fake accounts to sow distrust and social division in many parts of the world, in some cases fomenting violence against groups or individuals.

"These systems use a combination of signals such as patterns of using suspicious email addresses, suspicious actions, or other signals previously associated with other fake accounts we've removed," said Facebook analytics vice president Alex Schultz in a blog post.

The figures from Facebook's transparency report suggests Facebook is acting aggressively on fake accounts, said Onur Varol, a postdoctoral researcher at the Center for Complex Network Research at Northeastern University.

Many users can't tell the difference between a real and fake account, researchers say. Facebook and Twitter have been stepping up efforts to identify and weed out bogus accounts, and some public tools like Botometer developed by Varol and other researchers can help determine the likelihood of fake Twitter accounts and followers.

"A lot of the harmful content we see, including misinformation, are in fact commercially motivated," Zuckerberg told reporters. "So one of the best tactics is removing incentives to create fake accounts upstream, which limits content made downstream."

Having an army of realistic fake profiles at your fingertips gifts the owner a great deal of power. As social media thrives on herd mentality, the ability to artificially inflate the perceived support for a cause or topic has a dramatic impact on the world and can be auctioned off for a hefty price.

The same trend can be seen in the way Facebook reports on removing bots. On the one hand, the huge number of fake accounts which Facebook is regularly removing could suggest that they are on top of the situation.

Furthermore, we protect your ads from being seen by bots and other nefarious traffic sources, provide you with full transparency into every click, and even use cross-platform data to inform your other campaigns.

Set up in November 2012, the fake account had amassed almost a million followers by November 2018, when it was finally suspended by Twitter for impersonating the Russian leader. Confusingly, Putin's official, verified English-language Twitter feed has a similar number of followers.

The precise number of fake Twitter accounts is unknown. According to one Twitter staffer, however, each week the platform challenges between 8,5 and 10 million bots, with two-thirds of malicious accounts automatically removed. Facebook estimates that 5% of its worldwide monthly users are fakes. The social network deleted some 1.7 billion fraudulent accounts in the second quarter of 2021 alone.

Twitter, meanwhile, reserves the right to permanently suspend accounts for impersonating individuals, brands or organizations in a misleading or deceptive manner. Accounts that happen to have a similar username or profile picture to others are not, meanwhile, in automatic violation of this policy.

Genuine Facebook, Twitter and Instagram accounts operated by companies and persons of public interest usually bear a blue verification symbol. You will see this symbol on the profile pages of politicians and celebrities.

Take, for instance, this Facebook account supposedly owned by Elon Musk (see screenshot below). Registered to the misspelled username "Muskk" and "facebook.com/elonreeve.musk.338," it is highly unlikely that this is Elon Musk's actual Facebook page. Beyond that, organizations and persons of public interest will usually link to their real social media accounts.

If you want to verify the authenticity of an account registered to a person who is not in the public eye, use search engines to check whether said person is present on other social networks using the same name. Check if the profile pictures used are similar. Also check if profile biographies, contact and location details match up. And examine whether the accounts share similar content. If you can detect a large degree of overlap, then you are most likely dealing with a genuine account by a real person. It is very unlikely, after all, that fake accounts with identically named profiles and near-identical content are operated on different social networks.

You can also study profile pictures for hints (to learn more about spotting manipulated images, see the explainer in this series), provided one is available. Use the reverse picture search method (a service offered by Google, Bing and Yandex) to find out whether the picture used in the profile depicts someone else than claimed, or if the image has surfaced elsewhere online. Low-resolution pictures can be a red flag as well: it is very unlikely persons of public interest will use grainy pictures for their official social media profiles.

Is it likely Germany's former Chancellor Angela Merkel has fewer than 4,000 Twitter followers? And how plausible is it for the Duchess of Cambridge, Kate Middleton, to have a mere 52 friends on Facebook? Neither account seems particularly authentic.

Another way to assess the authenticity of an account supposedly run by a famous person is to check if verified accounts interact with it. Do teammates from the same sports club, for example, comment on photos shared by dubious account? Or do fellow, verified accounts repost content shared by the account in question?

Pay attention to when a social media profile was created. If it was set up and has been active for years, it could well be real. Still, this is no surefire way to gauge authenticity. The fake Putin account discussed above, after all, was active for six long years.

Be wary of people sharing content they seem to have little or no connection to. A German student posting images purporting to show an Afghan war zone? A Russian pensioner sharing photos from what she claims are Parisian anti-vaccination protests? Neither case seems particularly plausible. Both could well be fake accounts, or at least be sharing unverified misinformation.

Bots, short for robots, tirelessly comment on Facebook posts, share content, or artificially stoke online debate on marginal or otherwise overlooked topics. Their behavior, as the name suggested, resembles that of automated robots.

Bad bots, in contrast, are designed to imitate genuine human activity to advance a certain agenda. Depending on the algorithm, such computer programs may compose and publish social media posts or comments, follow others, or even send out friendship requests.

Bad bots can distort reality by amplifying certain political opinions on social media platforms, and drawing artificial attention to certain issues by repeatedly sharing falsehoods or disinformation and undermining constructive online debate.

Recently, researchers from Carnegie Mellon University analyzed over 200 million Tweets sent out in 2020 discussing coronavirus or COVID-19. They came to a worrying conclusion: 82% of the top 50 influential retweeters were bots; likewise 62% of the top 1,000 retweeters.

Have you observed a single social media account simultaneously publishing identical content across different platforms, or underneath several different posts? Does the account post countless crude replies in a very short time, or constantly retweet content? If the answer is yes, you are most likely dealing with a bot.

Bots are also known to typically follow a large number of other accounts, while having few to no followers of their own. To check follower numbers, consider using tools like Followerwonk or Botometer.

Trolls are real human beings who exhibit destructive and hyperactive online behavior, much like bots. They are often paid to harass certain public figures or media organizations. Facebook describes such targeted action as "coordinated inauthentic behavior on behalf of a foreign or government actor." Such campaigns can even be waged by so-called troll factories (to learn more about trolls and their role as agents of state propaganda, see the next article in this series).

One of the most prominent examples of a troll factory is the Internet Research Agency, in St. Petersburg, Russia. EUvsDisinfo, a project by the European External Action Service to counter disinformation, found that the agency had disseminated falsehoods and pro-Kremlin messages in multiple languages during various European and US election campaigns, and in the lead up to several referenda.

The agency gained global notoriety when US intelligence services published a report in January 2017 analyzing how it had worked to manipulate US public opinion ahead of the 2016 US presidential election. Since then, Facebook has been busily deleting accounts affiliated with the agency.

Troll factories and so-called troll armies are also said to operate out of India, China, Saudi Arabia and Mexico. Indeed, in 2020, The Washington Post revealed that teenage supporters of former US President Donald Trump were paid to spread disinformation on social media.

Troll accounts are harder to detect than downright fake accounts, as they are usually controlled by real human beings. Spotting troll accounts becomes even more challenging when they have been set up and maintained over years as part of wider troll networks. You cannot, therefore, easily identify them by their dubious profile names, recent registration dates or suspicious follower numbers. Even so, be sure to check for such potential clues.

In addition, closely study the content shared by the account in question. Does it contain links to websites known for publishing disinformation? (For more on detecting disinformation, see the first article in this series.) Pay attention to whether the user also publishes personal posts alongside sharing other content. If an account only reposts third-party content, this could indicate you are dealing with a troll. Be especially suspicious if this shared content contains disinformation.

3a8082e126