Twitter is closing in on a billion user accounts but as we looked at last month, nearly half of those accounts have never published a single tweet. And even if an account is active, that doesn’t mean it’s legit as a large number of these accounts are operated by social bots.
Plenty of automated accounts are easy to spot as they simply regurgitate news from a particular website or publish stock information. Others, however, are a bit more manipulative in nature and post questionable content such as fake news, rumors, spam, slander and the like. That said, how can you tell if a particular account is operated by a human or a computer?
Up to this point, the best solution was to simply use good judgment. Naturally, that doesn’t always work which is why a tool like the new Bot or Not app could be infinitely useful. Developed by researchers at Indiana University, the app takes a Twitter user handle and analyzes it against more than a thousand different criteria to determine the likelihood of it being run by a bot.
The app then gives a ranking on a 100-point scale based on patterns such as length of time between posts, the structure of the accounts’ networks and the content it posts.
Bot or Not is described as a work in progress. The team says they will continue to work to refine its algorithm as spammers will no doubt adjust their methods accordingly to circumvent the app’s analysis.