Fake friends: Social media accounts that are bots

A recent study from the University of Southern California and Indiana University, found fake Twitter friends to the tune of roughly 48 million active bot accounts, or 15 percent of reported active users that are not human by any definition.

Earlier this year, a computer scientist in London exposed expansive networks of fake Twitter accounts – with the largest consisting of over 350 000 profiles. The BBC reported, UK researchers accidentally uncovered the lurking networks while probing Twitter.

The fake friends may have been used to swell the numbers of followers, send spam, and boost interest in trending topics. On Twitter, bots are accounts that are run remotely by someone who automates the messages they send as well as running certain activities.

Some real users even pay for bots to follow their account or to dilute chatter about controversial subjects.

CNBC reported much bigger big chunks of those “likes,” “retweets,” and “followers” lighting up Twitter accounts that may not be as a result of human intervention.

Researchers at USC designed more than one thousand features to identify bot accounts on Twitter, in categories including friends, tweet content and sentiment, and time between tweets. Using that framework, researchers wrote that “our estimates suggest that between 9 percent and 15 percent of active Twitter accounts are bots.”

Twitter currently boasts around 319 million monthly active users, a figure that translates into some 48 million bot accounts, if USC’s high-end estimate is applied. The report also considered complex bots showing up as humans in their model, “making even the 15 percent figure a conservative estimate”.

At 15 percent, their evaluation is far greater than Twitter’s own estimates. In a filing with the SEC last month, Twitter declared that only up to 8.5 percent of all active accounts contacted Twitter’s servers “…without any discernable additional user-initiated action”.

The research could be troubling news for Twitter – which has struggled to grow its user base – if there are roughly 20 million more bot accounts than Twitter’s own assessment.

A Twitter spokesman told CNBC that some bots in the study’s calculation “are extremely beneficial, like those that automatically alert people of natural disasters… or from customer service points of view,” because they help shape user experience.

The study however highlighted the malicious applications of bots, like spurring fake grassroots political support.

But bot traffic is not unique to Twitter, as Facebook and Google’s “good bots” make up 8 percent of all internet traffic. The real concern is whether audience measurement companies should take bots into consideration as part of user traffic numbers, which affect advertising potential.