Fake Users and Fake Followers
The Problem
The technology section of today's New York Times has a feature-length article titled The Follower Factory. It details the black market for fake users and fake followers on Facebook and Twitter. People who make money by having lots of followers (wanna be celebrities, brands, authors) can buy fake followers from companies that create fake users of varying quality and will have those fake users follow the buyer's page or feeds.
So, if I wanted to make this blog seem popular and influential, I'd simply buy a few tens of thousands of fake followers, for a penny or two each.
Why should we care? Beyond the obvious fraud perpetrated on businesses like advertising that pay based on numbers of followers, there are two reasons:
Fake users are impacting our political discourse: It turns out that people give more credibility to posts and tweets from people and organizations with many followers. But, of course, there's no easy way to know whether those followers are real or purchased fake followers.
Fake users can affect real people's reputations: So-called high-quality fake users mimic real users, copying key aspects of their profiles like name, location, school, age, gender, etc. When fake users promote, for example, pornographic web sites, the reputations of the corresponding real users are affected. I have a (real-life) friend who has been plagued by fake users on Facebook that mimic him and then try to befriend his Facebook friends.
Towards Solving the Problem
This problem could be solved, or at least reduced, by Facebook (and the other social media platforms) if they wanted to do so. The approach is obvious: Provide some way to have verified users and a way to filter out non-verified users.
Verified Users
Facebook already has taken a step in this direction with their verified pages and verified profiles. Facebook verifies that profiles of certain celebrities, media companies, and brands are real and indicates so by displaying a blue badge (). This mechanism could be extended by providing a way to verify ordinary users like me and you, and by (optionally) filtering out the impact of non-verified users. For example, when looking at a purported news item, I might specify that I only want to see likes from verified users.
I don't mean to imply that this would be trivial do. But there are some shoulders to stand on. The US financial and legal systems partially solve this problem through the use of notary publics. If I want to mail a signed document to a bank, the bank might require me to sign the document in front of a notary public, who looks at my government-provided identification and certifies that I'm the one who actually signed the document. Social media platforms could build on this to allow ordinary people to verify themselves in this way.
Yes, this would cost somebody real money. Notary publics charge a fee for their service. Facebook (or whatever) would need substantial staff at least for a while to process the verifications, although one could imagine some level of automation over time. And, it is not clear how well the notary public infrastructure would scale if there were all of the sudden new needs for their services.
Who would pay for all of this? I'm not sure, but there are various models that could be tried. I think that the social media networks are soon going to get political pressure to address this problem, so there is value to them to spend money to get ahead of regulations that may be imposed. Advertisers might be willing to pay more for impressions of verified users. And, some users, might be willing to pay both time and money to become a verified user. Personally, I would certainly be willing to spend the time and money to have a verification form notarized and then mail it to Facebook.
A verification system like this would completely disrupt the business model of the peddlers of fake followers by raising the cost of creating a fake user from essentially zero to at least several dollars per fake user.
Filtering
The other end of a solution for the fake users and fake followers problem is a set of capabilities for filtering out their influence. Figuring out the best mechanisms would require experimentation. It could start with simply showing some data. For example, Facebook and Twitter could show both the total number of "likes" and the number of "likes" by verified users. Over time, as more people get verified profiles, there could be an option to filter out any impact from non-verified users.
Disenfranchisement
One of the democratizing aspects of social media is that it gives voice to anyone with access to the Internet. A verified user system like I've described would disenfranchise some people for exactly the same reasons that voter id laws in some states disenfranchise certain people. This would be an unintentional but negative side effect. Additionally, political protesters who count on social media as an organizing tool may be unable to take the risk of becoming verified users. One possible solution would be a notion of verified but anonymous user, in which Facebook would verify that there's a real person but would not publish real identification.