How common are fake Twitter accounts

10 questions you can use to recognize a fake Twitter account

Twitter has a reputation for providing the world's bot armies with a perfect habitat. Although the social network has done a lot in recent years to promote a “healthy culture of conversation”, the problem has not gone away. If you use Twitter, you will surely notice this again and again: You are followed by accounts where you can neither explain how they found you nor what connections could exist between you. Many women around me know, for example, the phenomenon that American (ex) soldiers briefly follow them and then disappear again. Are these all bots?

How manipulation is done on the internet and what consequences this has for our coexistence, our society and our democratic processes has been of particular interest to me since it has become clear that this puts democratic elections at risk. I have dealt with this topic extensively in this series on election manipulation using the example of Brexit.

In episode 1 of this series on social media manipulation, you've read a lot about how difficult it can be to identify fake accounts and why that is. Twitter is notorious for its “bot armies”. But as you've seen, real bots make up only a small part of the problem. There are an incredible number of bots that often only have a short half-life and are therefore so difficult to arrest. But they are only part of managed disinformation campaigns. For normal Twitter users like you and me, it is difficult to see behind the architecture and activities of these controlled actions. For this we need internet researchers and analysts. But you can use this checklist to gain more clarity if an account seems suspicious to you.

Every single point of the checklist is not in itself a clear indication that it is an automatically or semi-automatically controlled profile. But the more signs from this list you find on a Twitter account, the more likely it is that it is a fake account. With the following questions you can better arm yourself against the Trollbot and fake account problem.

1. When was the account created?

Noticeable are young accounts with a low number of followers who become aware of you for no apparent reason. If there are no mutual acquaintances and you cannot find a plausible answer to the question “Why is the account following me?”, Recently created accounts may be suspicious.

2. When was the last time the account tweeted?

It is noticeable if the account has not tweeted anything for a long time, but still starts to follow you.

3. How many grammar and spelling errors are in the tweets?

If an account has difficulties with the German language, that doesn't mean anything: Neither does it allow the judgment that the person belonging to it is stupid (dyslexia, for example, says nothing about intelligence), nor that it is a Russian bot that only has taken a German crash course. However, poor German or English can be an indication that the account is operated from abroad and belongs to a controlled campaign.

4. Is the profile photo real?

Many Trollbots use stock photos for the profile or photos of hacked accounts. You can use Google's reverse image search to see if the photo is being used for multiple accounts. If so, that is highly suspicious.

5. Is there a profile photo at all?

A missing profile photo can indicate that the account was created quickly. Bot hobbyists can also be lazy.

6. Does the account only tweet controversial and highly political content?

If you have the impression that all messages are intended to sow discord, provoke and defame, or if only extreme political positions are represented, the account is suspect. Do you know real people who would act like this? They are extremely rare and likely to have no friends. Why should you deal with such a profile online?

7. How often does the account tweet?

This question is less helpful than it seems. Because not all accounts that twitter a lot are suspicious. People who tweet from conferences, for example, want to draw attention to their topic, but do not immediately make themselves suspicious. A few years ago internet researchers set the magical limit of 50 tweets per day. However, that is now completely out of date. Bots can tweet a lot or a little, you can't draw a general line for this.

8. When does the account tweet?

An account that pretends to be German but tweets from a completely different time zone is suspicious. You can find this tool on the Scot or Bot website that shows the time zone of the tweets.

9. Which location does the account indicate?

Foreign fake accounts often use rather unspecific location information, for example “Germany” or “Europe”. Most real people on Twitter are much more specific and write where they live.

10. Has the account recently changed its name, bio, or agenda?

Fake accounts with many followers sometimes suddenly exchange all of this with the aim of addressing a different target group. Even if masses of tweets are deleted from your own profile, this indicates that a fake account that has been introduced is changing its function. Of course, real people change their positions from time to time, but they don't necessarily delete tons of old tweets, change their name, bio and agenda in one go.

What can you do when you have made a “wrong friend” or are attacked yourself?

If you think you are dealing with a bot or trollbot on a suspicious account, you should let Twitter know. Twitter describes how to do this here. If you are unsure whether an automated account violates Twitter's own rules, you can check it on Twitter. The Federal Ministry of Justice and Consumer Protection has information on its website about what you can do if you think that Twitter is not handling your report correctly, for example if the Trollbot distributes hate postings that could be criminally relevant.

If you have come under attack from a trollbot horde yourself, the recommendation of experts is not to block the accounts, but to mute them - i.e. to mute them. This makes it less obvious for the Trollbots that you are reacting to their provocation because they are not notified of it by Twitter. You should also take screenshots of the postings immediately, including the meta information displayed under the tweet such as date and location.

If that's not enough, because real people are already getting involved in the attack instigated by fake accounts, you can go on the offensive and ask your network for help. It is best not to publicly call for it yourself, but to ask someone to do it for you. On the one hand, it is important to communicate respectfully and firmly, and on the other hand, to gain as many supporters as possible in a short period of time. You can find further useful information in the report “Hate at the push of a button”. At the same time, the platform operators should be informed and asked to take measures against the attacks. You can also activate the internet guard of the responsible state police.

What Twitter does against fake accounts

In October 2018, Twitter published all accounts and their tweets for which it sees a connection to controlled disinformation campaigns from 2016 onwards. This is intended to help government and international defense organizations and research groups better understand how these disinformation campaigns are structured in order to be able to develop strategies for dealing with them.

Twitter is also now regulating the interface for app developers and Twitter analysts, the so-called API, to better understand which data is leaving the platform. However, only a telephone number is required, so it is not clear how much this measure prevents abuse.

In July 2018, Twitter also deleted 70 million accounts from the platform, which also led to some prominent users losing many followers. And probably also to the fact that Twitter recently had 9 million fewer daily users on the site. Still, Twitter says it wants to keep cleaning up. However, none of this will solve the problem, pretty much all experts agree on that.

When it comes to tweeting, the main problem is speed. That's why a lot is gained if you slow down and check the links that you distribute beforehand.


Click here for episode 3 of my series: 7 questions with which you can recognize a fake profile on Facebook.

Disinformation campaigns and misinformation are also a major problem with medical and health issues.

Editor: Theresa Bäuerlein; Final editing: Vera Fröhlich; Photo editor: Martin Gommel.