• Mark R Johnson

How To Spot A Bot

Updated: Jun 2


Bots have hit the news multiple times for their use in manipulating public opinion, influencing elections, and otherwise furthering agendas by spreading misinformation and artificially amplifying opinions.

Bots are estimated to make up approximately 10-20% of accounts and activity on twitter, but are estimated to comprise over half of the activity for some hot and/or controversial topics.

Below are some guidelines to help identify bot accounts. This article will primarily focus on Twitter, but much of this advice can be applied to other platforms as well.

Creation Date: Bots often have a fairly recent creation date as they are typically accounts with a particular goal of amplifying certain types of material. However, this is not always the case. Bots can be used for multiple related campaigns or for amplifying the same type of material over a longer period of time. Additionally, some bots will have older creation dates, but activity only dating back to their current campaign or mission. Aged accounts, particularly on Facebook, are a common method of both circumventing security and bot detection as well as adding legitimacy to the account.

In fact, there is an entire online market dedicated to selling aged accounts with different types of established profiles and/or followers.

Usernames: Particularly on Twitter, many bot usernames or handles will have long strings of numbers in them. They may also have usernames and handles that are unrelated to each other.

Biography and personal information: Bot accounts often have no biography or personal information. However, others will have information or phrases that match their agenda and function.

Photo: Many bots have no photo at all. However, many others will take photos from real individuals and modify them slightly. This tactic is particularly common on Facebook where photos are now necessary for creating new accounts.

Content: Some bots will primarily retweet content rather than posting their own. Others will post a high volume of tweets with the same hashtag or set of hashtags. However, not all are quite this obvious. Some will not necessarily use hashtags, but their language will be stilted or somewhat inappropriate for the tweet they are responding to. Others, when looking at their history, are replying to tweets in a fashion that is obviously triggered by certain words or phrases in the original tweet. Indeed, the bulk, if not entirety of their activity, will be replies.

Posting Frequency: Many bots post an incredibly high volume of tweets - more than even the most internet-addicted human being could feasibly post. Others that don’t post quite that often might still post at overly regular intervals and at the same times every day, even if they have been programmed to more accurately mimic human activity and sleep schedules.

Follower count: Bots will frequently have a relatively low follower count, often with none at all. Conversely, bots might have an incredibly high number of followers and accounts that they are following, but further inspection will reveal that many of those accounts are, themselves, bots.

Networks: Relatedly, many bots work together in tandem for the same cause they will often all follow each other and/or post the same content around the same time. Some will also reply to each other - either supporting or “fighting” each other, creating an artificial conflict even though both accounts are controlled by the same individual or organization. In fact, false conflicts created by bots and sock puppets are common means of manipulating a narrative or drawing additional attention to a cause in a less direct way.

It’s important to note that threat actors are aware of these tells and will, depending on the sophistication of their operation, attempt to circumvent them. Not all bots will fit all of the above criteria. Indeed, not all “bots” are fully bots in the traditional sense, but more of a cyborg - partially automated, partially human-run account. However, these tips and tricks will still be effective in helping you identify bots and fake accounts and stop the spread of fake news.

#fakenews #bots #howto