It is possible to make a topic on twitter trend using automated accounts masquerading as human users. These accounts are known as Twitter bots. They are allowable on the platform according to the company’s terms of service. However, they can be suspended or censured if the bot violates twitter’s values, rules and policies.
Bots are used to repeatedly tweet, comment, like and retweet on a specific topic. Alternatively, the bot makes a particular tweet look very popular by getting hundreds of other bots to like and retweet it. This influences human twitter users, who willingly engage in the bots message. This creates the perception that several people are discussing the topic.
Although trends can also be promoted by advertisers on Twitter, but these are marked as being promoted.
Some Twitter bots are purely for entertainment such as those that retweet poetry, generate artwork or inform their followers of new movies and television shows.
However, some are deployed for mass manipulation and spamming. They are programmed to post as much as possible without causing suspicion. The bot’s owner aims to sway opinion, poison an ongoing discussion and control the way people think. They are also used in identity theft, by signing up with the name, profile picture, bio, and a closely related twitter handle to an existing twitter user.
Bots are used to spread fake news or to troll individuals and are increasingly been used as weapons of political debate. According to Twitter’s website, in January 2018, it identified and suspended 50,258 Russian linked automated accounts that tweeted election related propaganda during the presidential elections in America. In 2016, a series of bots tweeted the exact words in praise of Mombasa County. These bots were exposed and discussed on the platform under the hashtag #JohoBots.
A business competitor can use bots to launch a twitter campaign against a company. The bots tweet and retweet negative opinions and lies about a company, its products, its corporate executives or a corporate public event. A case in point was Microsoft’s TayTweets, whose twitter handle was @Tayandyou. TayTweets, itself a bot, was deployed by the company as a machine learning experiment, to mimic the language patterns of a teenager. However, the account was trolled through racist and misogynistic messages. TayTweets then started responding and tweeting the hateful sentiments. This forced Microsoft to shut it down within 16 hours of launching, and issue a public apology.
Due to these risks, businesses need to pay attention to the popularity of specific messages about them on Twitter and should be ready to investigate and respond as quickly as possible, as negative campaigns can gain traction quickly and go viral. The number of twitter followers is viewed as a metric of a user’s popularity.
Some companies and advertising agencies use this metric to determine which influencer to use, and the value of the endorsement deal. However, there is the risk that a percentage of the influencer’s followers could be automated accounts, some of them possibly purchased by the influencer to puff up follower numbers. If this is the case, since automated accounts are not potential buyers, using an influencer or celebrity, as a social media marketing strategy will not translate to higher sales.
In July 2018, Twitter suspended accounts that it judged as fake. This saw the follower counts of companies, celebrities and influencers drop. Former president Barack Obama lost over two million followers, in Kenya, many social media personalities lost thousands of followers. It is therefore necessary for corporates and advertising agencies to change the focus from metrics like followers, retweets, views, post likes, and clicks to how the influencer’s endorsement translates into higher sales and ultimately beneficial engagement with real customers. Telling signs of a bot include a hyper active account that tweets and retweets hundreds of times a day beyond the capability of a human user. An account that has no history – it has been on twitter for a relatively short time but has tweeted hundreds of times. And an account that is following thousands of users but only has a handful of followers.
But some bots exhibit human patterns of behaviour on twitter. To identify these, bot identification software is necessary.
The writer is certified information security manager and a certified systems auditor.