Massive networks of fake accounts found on Twitter

This seems like a pretty straightforward thing to identify. I remember when the NYT's story, "The Agency", was published [0], some of the fake accounts it had mentioned were still up. Even though the accounts in that story were actually populated by real people, the sockpuppetry was pretty easy to identify. One: the accounts' past tweets right up until they started spreading the news about the fake U.S. disaster were in Russian. Two, all of the tweets of the fake news had almost exactly the same number of favorites and retweets (around 300), and you could see that everyone in that cluster was just retweeting each other.

I'm more fascinated by the spam by Facebook accounts. These show up all the time in relatively popular comment sections, and yet apparently FB doesn't care, or the problem is trickier to automatically flag. For example, this comment [1] is clearly spam...but if you click through to the account, it seems to be a real person [1], with a normal-seemingly friend network, mundane photos of life that aren't obviously stock photography. There are a few junk comments (a bunch of "hi's", but as an outsider, this is what makes FB a lot trickier to analyze, because you don't know how much privacy that user has enabled on their own account.

[0] https://www.nytimes.com/2015/06/07/magazine/the-agency.html

[1] http://imgur.com/a/Rr8d3

[2] https://www.facebook.com/gulfam.raj

> A Twitter spokesman said the social network had clear policy on automation that was "strictly enforced".

> Users were barred from writing programs that automatically followed or unfollowed accounts or which "favourited" tweets in bulk, he said.

I am constantly getting followed by accounts with tens or hundreds of thousands of follows and followers, usually checkmarked accounts though I've never heard of them. It's painfully obvious these verified users are using bots to randomly follow people, both to spam my inbox with "you have a new follower" messages and to encourage people to "follow back".

But Twitter does nothing about it. It's not "strictly enforced" at all.

I think this story is getting more coverage than the one from a few months ago that showed that, for instance, both Clinton and Trump had millions of fake accounts following them:

> Per eZanga, 4.3 million—or 39 percent—of Trump's more than 11 million Twitter followers as of August came from fake accounts while the other 6.7 million are actually real users. And for Clinton, 3.1 million—or 37 percent—of her more than 8 million followers were fake while 5.3 million come from real accounts.

http://www.adweek.com/news/technology/more-third-presidentia...

350,000 is about 0.1% of Twitter's user base. Does anyone here think the number of fake accounts isn't orders of magnitude higher than that?

Bot accounts, not fake accounts. I don't even know what the latter means—plenty of people don't associate twitter with their real name. And why would you!

Secondly, of course there are this many. There are probably many more. I run several bots myself; there's nothing wrong with this.

Twitter's TOS is only as good as its enforcement, and if there's anything twitter is terrible at, it's having any control over its community.

In their recent SEC filings, Twitter estimates around 5% of the MAUs are spammers or bots. They also estimate they have 317 million MAUs, which when you work it out gives around 16.7 million monthly spammers or bots.

Numbers come from their 2016 Q3 filing https://investor.twitterinc.com/secfiling.cfm?filingID=15645...