Twittering Bots

The use of Twitter, Facebook and other social networking sites by people involved in the recent unrest and overthrow of governments in the Middle East shows that this technology can be a real force for social change.

However, this power makes it very attractive for advertisers, lobby groups and governments to try and steer individuals and social groups into certain types of behavior.

Celebrity Tweets For Money

Many celebrities are secretly paid to 'tweet' about a certain product in the expectation that their Twitter followers will want to be 'on trend' and buy the product. Lobby groups may want to 'talk up' a point of view, and incumbent politicians can get their supporters to flood 'talkbacks', using social media. However, sending messages one at a time to specific individuals takes time and effort. It's much easier to use social bots to send bulk messages just the way spam is sent - automatically.

Commercial Uses

Because people put so much information about themselves online it makes it very easy for commercial companies to use technology to trawl the internet to discover your online search and purchasing behavior. Once they have gathered the information they can use social bots to manipulate your buying patterns in their favor. This type of technology can access thousands of people at one time and is very cheap to run. It means that corporations can target you with very direct networking campaigns designed to appeal to you more specifically than ever before at an extremely low cost. And you won't even be aware that you are being manipulated in this way.

Infiltrating Networks

A recent experiment showed that a 'social bot' can 'tweet' messages and infiltrate social networks without people being aware that they are 'talking' to a robot. Although the technology doesn't meet the Turing test yet, it is still very effective in influencing social behavior. In the experiment, the Web Ecology Project organized a two week competition to create 'social bots' that could imitate conversations on Twitter. The idea was to see if they could get other Twitter users to 'follow' them. The three participating teams were able to infiltrate a network of 500 Twitter users and get them to 'follow' their bots. One bot got over 100 followers in the two week period without being detected. According to the organizers the bots were "able to heavily shape and distort the structure of the network"

How Can You Tell It's A Bot?

Twitter comments are kept deliberately short which makes it easy for social bots to interact as if they are human. According to Tim Hwang, the man behind the Web Ecology Experiment, if you suspect someone you are following on Twitter may not be who they appear to be, try having a longer 'conversation'. If they don't seem to respond in a human or intelligent way you may be 'talking' to a bot. Hwang has a new large-scale social architecture experiment where he hopes to use social bots to create a bridge to connect two large previously unconnected Twitter groups together. This will involve 10,000 unsuspecting people having 'relationships' with robots.


The implications of this experiment are wide-ranging from positive ones like encouraging support for humanitarian causes or social participation in events to negative ones like being manipulated by governments to conform to certain behavior. The US military is already looking at using this technology in Afghanistan and Iraq to create fake personas to help counter enemy propaganda and violent extremist behavior.

This may be good against terrorists, but what if for example governments use bots to influence election results in their favor?

Log in or sign up to comment.

Post a comment

Log in or sign up to comment.
86% of email addresses posted on websites are used by spammers to send unsolicited emails?

63% of all "Remove me from your list" requests are not honored.

Spam accounted for 80% of all e-mail received in 2004, up from 62% in 2003