Researchers pose as scammers to curb spam on Twitter

Any regular user of Twitter will testify against how annoying spam and spammy accounts on the micro-blogging website can get. However, there are a bunch of people who’re planning to be your knights in shining armour and find a way to curb this spam attack on Twitter. You know how they say, when in Rome do as the Romans do, right? These people are doing just the same – being a part of the spam to stop the spam.

A research team consisting of participants from George Mason University, the University of California, Berkley and the International Computer Science Institute has teamed up with Twitter to study exactly where these accounts come from, how they work and how they can be stopped. To achieve this goal, the team purchased about 127,000 fake accounts – automatically generated – from 27 different merchants between June 2012 and February 2013.

More from News & Analysis

What is the US HIRE Bill and why is India’s $250-billion IT sector worried? Is the internet dead? What's this theory that OpenAI's Sam Altman says might be true?

Curbing spam, one fake account at a time (Image credit: Leihut.com)

There were some quick findings as soon as the research began. Merchants employed countermeasures to try and circumvent technical stop signs put up by Twitter. First up, in order to set up an account on Twitter, you need an email ID and you need to be able to solve CAPTCHA image puzzles during the sign up process. The paper presented by the team elaborates on how easy and cheap it is to get auto-created email IDs on services such as Hotmail, Yahoo! and Mail.ru. As far as solving annoying CAPTCHAs and being able to provide IP addresses for these fraudulent accounts, India ranks high on the list of places to be to get yourself spam accounts.

The team, after a long period of study, was able to set up some flags that could work as identifiers while searching for auto-generated accounts. These indicators include a specific pattern of usernames. The results ended up helping out Twitter massively, for after weeks of being provided with the data, Twitter managed to cull “several millions” of accounts that were auto-generated by those 27 merchants. When the team tried to buy 14,000 more accounts from these very merchants, 90 percent of accounts came to them dead. A merchant told the researchers, “All of the stock got suspended… Not just mine… Don’t know what Twitter has done.”

Rejoicing would be a little too premature at this juncture as far as saying goodbye to spammy accounts go. Twitter will have to strive hard in order to be one step ahead of these merchants who produce fraudulent accounts all the time. The merchants have figured ways to beat the system by having sweatshop workers enter CAPTCHAs and using fake IP addresses to register accounts while Twitter is unable to catch hold of all of them. The struggle to stop spammers on social networking websites seems eternal, but Twitter it seems has found a way to curb part of it thanks to these researchers.