Fake Followers for Hire, and How to Spot Them

A team of researchers at UCSB, led by Prof. Ben Zhao have explored a practice they've coined "crowdturfing" in which organizations create a false positive reputation on social networks such as Facebook, YouTube, or Twitter.  They've discovered that much of the activity on crowdsourcing websites involves asking workers to follow or like particular social network posts. In an effort to identify crowdturfing they have developed machine learning software which can detect crowdturfers on China's version of Twitter with 95 to 99 percent accuracy.

The full article, published in the MIT Technology Review, can be found here and the group's USENIX Security Symposium 2014 paper can be found here.