Twitter took decisive action this week by permanently suspending the account of Milo Yiannopoulos, a Breitbart.com technology writer, for inciting his followers on the social media service to electronically harass comedienne and actress Leslie Jones, who is currently starring in the all-female remake of the 1984 comedy sci-fi/horror movie Ghostbusters.
Yiannopoulos has been highly critical of the movie for the past several months up through its release. He's ravaged the film's script, plot, and the performances of its cast. As an opinion editorial columnist, he can criticize mass media and the entertainment industry, and nobody should object to that regardless of how abrasive his writing style comes across.
Twitter had to take action against Yiannopoulos because this high-profile incident forced it to reckon with its reputation as the social media site where angry mobs rule. But while its punishment of Yiannopoulos is completely justified, it doesn't get to the root of the problem.
Trolls can swarm your mentions and make all communication impossible. And the stuff they post can be truly vile: death threats, graphic images, and -- worst of all -- doxing. The recipients of such awful behavior are almost exclusively women. GamerGate, of which Yiannopoulos has been one of its most toxic voices, is a prime instigator of such misogynistic conduct.
Yes, there are a lot of trolls on Twitter. It is a highly-permissive environment and a technologically-straightforward service in which everyone is allowed entry and unrestricted access to anyone else participating. Everything happens in wide-open, town-crier fashion. It is a public square, with a massive amount of noise being generated.
To use information technology terms, Twitter is nothing more than a centralized messaging bus or a messaging queue (MQ). Yes, it is an extremely high-volume MQ, but it's an MQ just the same.
Twitter is a powerful tool precisely because it is a firehose of data. That is why it has value as the pulse of citizen news reporting. But, because it is unfiltered, users have no real tools for managing that flow of information, and a lot of that information is unwanted and inaccurate.
The opposite of Twitter would be something like Facebook -- where the information presented to the end-user is controlled by an algorithm, so end-users only see what the service believes actually interests them.
While that automation-enabled, echo-chamber approach works for Facebook, it wouldn't work for Twitter. It would completely devalue Twitter as a broadcast/message bus type of platform. But, at the same time, you can't allow throngs of hate-filled users to attack other people on the service.
One way is for Twitter to hire a nanny force and start penalizing high-profile trolls like Yiannopoulos. It could make examples out of them and send strong messages about how that sort of behavior will not be tolerated on the service.
But we're talking about a service with over 300 million users. That approach is only going to work against a small percentage of the population, so the only way to deal with this massive problem is to give the population the tools to police itself. And that's by providing the needed application programming interfaces (APIs) that could allow the community to produce more sophisticated means of filtering the wheat from the chaff than what comes out of the box.
Twitter has some basic controls. You can block users and report them as spammers. There is also the notion of protected versus public tweets -- but that diminishes the utility of the service. These controls are also reactive rather than pro-active. The key is to not receive unwanted communications in the first place.
A good example of how to best implement this is through community-maintained filter lists, such as those used in browser extensions, like Adblock Plus, and various email spam block lists used by service providers.
The web and email, however, are open-protocol standards, so it's much, much easier to implement these things. Twitter, while a wildly-successful platform, is a privately-owned and managed service.
Additionally, in recent years, Twitter has shut down a lot of the ability for third parties to tap into and extend its platform by reducing the amount of API calls third-party clients can make in addition to limiting API calls to server-side code.
What does any of this have to do with trolls? Everything. If the only valid Twitter clients are those made by Twitter, then Twitter has to do all the policing itself.
Not only does it have to waste energy on this now, but it also has the added baggage of appearing biased with its decision-making. Who is to say a first-class troll, such as Yiannopoulos, isn't as worthy for eviction from Twitter as a mainstream political candidate? On either side of the equation?
By restricting access to its APIs, Twitter has in effect created a closed environment, where trolls can operate freely, because end-users do not have access to tools that would enable them to insulate themselves. It means the bad guys and the angry mob can game the system a lot easier.
Time and time again we have seen that communities have a tendency to thrive when they have an ability to police themselves. By closing off developers in the last three years, Twitter has made it nearly impossible to do that, and it is now suffering the consequences as a result.
Does Twitter need to rethink its API strategy to enable users and communities to insulate themselves from abuse? Talk Back and Let Me Know.