Twitter's Fake Followers Watching IPO Closely

kraksmoka writes "Is your social media pro 'making it go viral' by pressing a button instead of interacting with a real audience? The purchase and use of fake followers by small to mid-sized social media agencies is rising on Twitter and there is concern that the growth of fake followers can't be stopped. "

Paid commentors

[globaljustin]

Virtually all the posters here appear genuine

browse at -1 and have a look at the comments...i mod often and you're right /. is *definitely* more genuine than most... /. is crawling with paid Public Relations staffers (Fox News is def. not the only one to do this), paid commentors, and maybe even an actual experimental bot (APK...)

They ruin the top of the comments on anything to do with Snowden, the oil industry, and the Trayvon Martin case type stuff....techies havent' gotten *more* conservative in the last 10 years...but /. comments on average have...it's because of PR and paid commentors

We *genuine* humans need to be more discerning than ever...there are people, much like us, whose entire job is to create false perceptions on things like /.

Its kind of important, for you know, idea neutrality that we all be smarter, respond to only comments that are value added and of course...and I need this advice as much as anyone...

***DONT FEED THE TROLLS***

Re:Paid commentors

[game kid]

A moderator a day keeps the Real Name policies away.

Re:A simple tech solution

[sinij]

>>>Twitter can determine when, where, and from what IP address an account is created.
 
I expect ./ crowd to at least understand that IP is not a reliable identifier. Twitter can only reliably determine when, everything else they know only if bot creator did not bother to spoof it. Behavior-based detection is also problematic - you can easily scrape existing activity, filter out swearing and specific identities, substitute location identifiers for something local and have 100% undetectable bot.

Example: Scrape small-town phone book, run permutation algorithm on second name and street # to avoid collision with real people (but keep everything else intact), add random gender-appropriate picture and follow a random set of big news and artists at creation. Pipe this through TOR, stagger your account creation to avoid tripping volume detection and mind timezones for posting and registering. Proceed to post random scraped tweets that are filtered for positive-biased sentiment.