Learning Search

Bots On Twitter: Why Are There So Many People-Imitating Bots?

I try – oh boy, I desperately try – to avoid politics in ResearchBuzz. But sometimes you come across a fascinating story or article idea, and there’s politics associated with it. For that I apologize in advance.

Chatter started circulating on Twitter this week about Donald Trump’s follower count on that network, his favorite for proselytizing. Newsweek published an article yesterday afternoon: Nearly Half Of Donald Trump’s Twitter Followers Are Fake Accounts And Bots. From that story:

“Because Donald Trump is the president of the United States and the most famous person on the planet, one wouldn’t think he would need to employ a bot to boost his Twitter following. It appears, however, he might have done just that. As screenwriter John Niven pointed out Tuesday morning, Trump’s Twitter account saw an unusal spike in followers over the weekend, many of which appear to have been created artificially.” Daily Dot wrote that the @RealDonaldTrump account had gained 900,000 of what it described as “fake or inactive” accounts in May.

(For the purposes of fairness, let me add that Twitter is pushing back on this story.)

Why this is happening is a matter of conjecture, but it does bring up the topic of bots on Twitter.

What are bots, why are there so many bots imitating people – and why should we be worried about that?

What’s a Twitter Bot?

Basically a Twitter bot is a computer program. And I’m defining it that ambiguously to give you an idea of what a broad scope of entities can be defined by the phrase “Twitter bot.” The most basic Twitter bots do things like auto-DM you when you follow them. Did you ever follow someone on Twitter and two seconds letter get a DM thanking you for following and then asking you to sign up for a newsletter or some such? That was probably a bot. Then you’ve got the bots that do things like watch for a single grammatical error with the sole purpose of correcting it. A little further up the scale there’s a bot that looks for and retweets accidental haiku. There are even Twitter bots that will glitch up and alter your images.

screenshot 2017 05 30 at 6 47 22 pm

These are all innocuous bots, possibly very clever but not particularly powerful. The bots that cause more concern are those which are designed to impersonate humans. Some of these are experiments – Microsoft had an AI bot named Tay which was yanked off Twitter after it went nuts – but many of them are designed to mimic human activity online without disclosure of what they are.

But why, you might ask, would a bunch of bots want to imitate people?

Because we’re social animals. More about that later.

Human-Imitating Bots for Social Media Credibility

You get a certain amount of cred on Twitter for how many followers you have. If you use Twitter, consider this: someone mentions you on Twitter so you look at their profile. Imagine they have 500 followers. Then imagine they have 50,000 followers. Aren’t you going to treat them a little differently if they have 50,000 as opposed to 500 followers? A large amount of followers is a social clue: this person is engaged and a lot of people want to know what they have to say.

… unfortunately it’s not necessarily a legitimate clue, as there are plenty of services which make money by selling fake follows. There are entire Web sites about how to buy Twitter followers. For as little as $6 per thousand you can buy your own fake followers to pad out your follower count and give yourself a little extra social media slick.

Just because someone has fake followers, though, doesn’t mean they were purchased. There’s a tool called TwitterAudit that analyzes up to 5000 of your followers to see if they’re real or, according to their analysis, fake.

Here’s TwitterAudit’s methodology, from its about page: “Each audit takes a sample of up to 5000 (or more, if you subscribe to Pro) Twitter followers for a user and calculates a score for each follower. This score is based on number of tweets, date of the last tweet, and ratio of followers to friends. We use these scores to determine whether any given user is real or fake.”

ResearchBuzz, according to this tool, has 96% real followers on Twitter.

screenshot 2017 05 30 at 7 35 01 pm

Okay, so people can buy fake bot accounts to make their Twitter followers look better. So what, right? What’s the big deal?

The big deal is what can happen when there are many, many Twitter accounts acting en masse.

Donald Trump’s Twitter account, according to Twitter Audit, has about half fake followers:

screenshot 2017 05 30 at 8 06 07 pm

And to maybe make this a little less political, that’s not the worst score I’ve seen on a large Twitter account. Katy Perry’s Twitter account is pretty abysmal according to Twitter Audit.

screenshot 2017 05 30 at 8 11 41 pm

But back to the issue at hand. Why should we care if Donald Trump’s Twitter account has 15 million fake followers?

Because we are social animals.

Bots and Propaganda

Twitter bots can not only pad a follower count, they can also push misinformation and propaganda through social media networks, and might go undetected. After all, you might not vet every single person who follows you, and while you might fact-check a tweet, how likely are you to vet the account which tweeted it?

This is not new news. Twitter bots have been working to spread propaganda and trying to shape public opinion since at least the first presidential debate last year. More recently, Heatstreet ran an article about French politician Marine Le Pen’s Twitter bot followers.

Obviously the bots cannot vote. But by spreading misinformation, or amplifying positive or negative messages about a candidate, they can encourage us humans to do the same thing – because we’re social animals. Propaganda bots might give an impression of popular support that doesn’t exist, or attempt to normalize certain perspectives by repeating them over and over.

As a matter of fact, the University of Georgia recently announced some research about this. In the article “Connecting the bots: Researchers uncover invisible influence on social media” you’ll find this quote:

“Although we’ve known about Twitter bots for years, the new research, recently published in Academy of Management Discoveries, marks the first time that bots’ social clout was studied in the field of information systems and management. Because of the increasing prevalence and sophistication of bots, their invisible influence may be affecting news reports and social media research, said Elena Karahanna, study co-author and professor of management information systems at Terry [College of Business].”

And that, in a nutshell, is why the fact that Donald Trump has so many bot or fake followers should cause concern. Not only can a critical mass of bots drown out human voices and human reactions, but it can also bend what news media decide to talk about and the way social media is assessed by academics.

It may be that Donald Trump and his administration have absolutely zero to do with his Twitter following and the number of bots it has. It may be sheer coincidence or the impact of being both politically- and Twitter- famous. But considering the current investigations into the election itself, the interference from bots that occurred during the US presidential election and others, and the research establishing what kind of damage organized Twitter bots can do, we have to ask: what is happening? What kind of due diligence are we (or Twitter) doing to make sure these bots are not being used to interfere in American governance? And how are we going to make sure it doesn’t happen again?

7 replies »

  1. Thank you, Tara, this was the first article I’ve read that made perfect sense to me on what bots are and what dangers they represent.

  2. I think this is another TPC public service.

    And in addition to the Putin Brothers and their ill doings, there comes to mind an odd question: Do bots have a right to freedom of speech? Right now this is very much a side-issue, but as AI’s grow in complexity…

  3. I was given a 10 out 10 rating for being a bot yesterday. Since I am a living breathing American voter, that is disturbing. In researching, I have found a spate of articles similar to this one all coming out the first part of June. I have also been approached by a reporter through Facebook trying to determine if I am, indeed, a real human being. From what I see so far, bot is now a new way to attempt to marginalize those of us who support President Trump. But I will keep investigating. I would rather have real followers so I will audit mine and remove “fake” accounts. But I will do so carefully since I know that I am not fake and yet I seem to fit the parameters.

Leave a Reply