As of quarter 4, 2016, Twitter had over 319 million active users across its micro-blogging platform.1 According to a study undertaken by The University of Southern California and Indiana University, 15% of all of those active accounts are bots.2 These accounts are not operated by an actual human behind the wheel, but a software tool with hidden ulterior motives. The motives range from spreading malware through compromised links, to spamming other Twitter users with affiliate links.
Other social media platforms have also been plagued with the problems of fake bot operated accounts, however, Twitter’s problems seem more evident than others. According to Facebook, 8.7 percent of their accounts are fake.3 Taking the total Facebook fake account tally to over 83 million. Instagram, another social networking website owned by Facebook, has over 8% of its total user base labelled as bot accounts.4 The paper titled ‘Online Human-Bot Interactions: Detection, Estimation, and Characterization’ highlighted a Twitter bot detection tool that makes use of over 1000 features that helps it classify a Twitter account as a bot. The bot detection tool analyzes the following metrics:5
- Detailed analysis of an accounts friends list
- Number of followers / following
- Broad analysis of tweets from an account
- Number of tweets pushed per hour
- Number of retweets pushed per hour
- Usage of emoticons
- Valence score of tweets, among others
The result of the tool? A staggering 48 million fake accounts were detected.
Twitter’s problem with fake users runs deep. According to the another fake user detection tool, 19% of Lebron James’s 34,234,037 Twitter followers are fake.6 23% of Kourtney Kardashian’s 21,871,706 followers are fake, and 16% of BBC’s 17,958,161 followers are deemed fake by the tool.7 These fake bots act more of less like real users as they have an algorithm governing their behavior. These bots can like, tweet, retweet, follow, un-follow and even message.
Problems of fake accounts
Fake bot based Twitter accounts harbor a multitude of problems. Some of the common issues that genuine users face are:
- Some bots spam affiliate links with the aim of generating revenue for its master
- Other bots impersonate other users and hijack their online identity
- Certain harmful bots also aim to spread malware throughout the social network, some with the aim of hijacking a user’s network through ransomware.
- These bots have the ability to influence news and trends across the community
Impact of fake accounts on Twitter’s bottom line
An internet company is largely valued with respect to the authenticity of user data it holds. The higher the legitimacy of data, the higher the value it commands. As a large chunk of Twitter users are fake or automated, Twitter’s valuation has suffered. The stock tumbled nearly 63% since its IPO in 2014.8 It listed at $41.65 (in 2014) and is currently at a valued at just a shade over $15 per share. Authentic data breeds legitimate and effective marketing opportunities, Wall Street recognizes this.
Helpful Twitter bots
Not all Twitter bots are created with the aim of spreading spam and adverts through Twitter’s network. There are a few automated Twitter bots that have actually helped the micro-blogging community with timely, precise and valuable information. A few helpful Twitter bots include:
SF QuakeBot: A bot that sources real time earthquake information through The United States Geological Survey and alerts Twitter users from the San Francisco Bay area.9 The bot has garnered an impressive 134 thousand followers.
Olivia Taters: The bot was created by coder Rob Dubbin. Olivia tweets about everything from getting late to work to watching her favorite television soaps.10 Though the bot is not particularly helpful to followers, its robotic yet seemingly human interactions have gained her over 5000 followers on the social networking site.
Pixel Sorter: The bot was created by Front-end developer and instructor @wayspurrchen. The bot uses one of many predefined image altering techniques to aesthetically alter any images tagged to the bot and spits out an artistic version of the image.11
Every word: This bot was created with the aim to slowly and steadily tweeting every word in the English dictionary. The bot has tweeted over 109 thousand times and amassed over 70 thousand followers while doing so.12
Google Facts: The bot tweets interesting and insightful facts and information on an hourly basis. Its excellent info has resulted in an impressive follower base of over 1 million.13
Dear Assistant: A user may ask any question and tag the bot. The bot will then answer that question and tag the person back.14