Oz Blog News Commentary

Botting Politics

January 19, 2017 - 07:05 -- Admin

During the 2012 US Presidential election, the Romney campaign was accused of buying fake Twitter followers from a bot network after he gained a whopping ~150,000 new followers in one day.  Romney's following jumped 17% in one day, providing him with a significant bump in the general perception of his grassroots popularity and support for a paltry ~$2,700 ($18 per thousand followers). 

RomneyDespite the accusation, the astroturfing effort worked.  Few saw, read, or cared the report that claimed these followers were bots and not people.  Since 2012, there's been a significant increase in the use of political bots in nearly every country in the world, from Afghanistan to Venezuela.  Most of these early efforts, usually done in support of the government or establishment, fall into the following categories:

  • Support padding (like Romney)
  • Disseminating or amplifying propaganda (Saudi Arabia does lots of this)
  • Demobilizing or trashing protest movements  (flooding hashtags for example)

Swarmed Mass 

Most of these early political bots are used in large swarms.  Swarms that allow them to act as a single unit to overwhelm the opponent with mass.  In some cases, the swarm is used to repeat a message again and again to increase its influence (retweeting, reposting, etc. en masse).  In others, it's used to flood an opponent with a large volume of responses refuting their claims.  

To really get a sense of this, let's look at a single Russian swarm (out of many) that was used to pump kompromat and demobilize the opposition inside Russia.  Botnet swarms like this flood any negative post, story, or tweet on a particular topic with pushback and amplify the dissemination of kompromat across social networks.  

After the shooting of Boris Nemtsov in 2015, this botnet sprang into action claiming that the Ukrainians shot him:

Botnet swarming effect

By using the breadcrumb trail of identical posts, Lawrence Alexander used social networking analysis tools to map out the swarm.  The result was a tightly interconnected network of 2,900 bots acting in concert to demobilize opposition and promote kompromat.


Big, crude swarms like this are already losing effectiveness as the social networking companies get smarter about detecting and banning them.  Naturally, this has created an arms race between bot makers and the social networking companies, but with a twist.  

The twist is that most governments aren't on the side of the social networking companies.  Most are working on bot networks that circumvent controls in order to influence and control political opinion just like everyone else.


John Robb


PS> Bots come in two flavors:  

  1. software or
  2. a combination of software and hardware (robots).  

While hardware based bots like drones have some scary/amazing (yet largely unexplored) tactical utility, most of the real action is in software bots or more specifically, social bots.  Social bots can be run from a single computer using multiple social networking accounts.  Others are operate as a network, using PCs compromised by malware.  In general, social bots can do the following:  

  • automate and amplify interaction with social networks.  
  • can converse with people (chatbots, customer service bots, etc.) -- some of these are getting amazing.
  • actively and remorselessly troll, harass, and confuse opponents.

Obviously, all of these attributes make social bots extremely useful politically.