Computing Blog

Now it’s making the national news, a few more thoughts on the Twitter troll issue.

Twitter’s problem isn’t individual trolls; they can be blocked easily enough. From what I can tell, the big problem is the large-scale pile-ons that overwhelm their target’s “Connect” tab. There is evidence to suggest these pile-ons are co-ordinated on other sites such as Reddit. It’s not just misogynist troglodytes attacking outspoken feminists, although that’s what’s getting the headlines; from recent evidence feral One Direction fans can be just as bad.

Which makes wonder if one possible solution would be to give users more control over what’s is and isn’t seen in their Connect tab. The default of seeing everything bar accounts you’ve actually blocked works for us ordinary folks; it lets people you’re not actually following join conversations and can be a way of discovering interesting new people. It’s easy enough to plonk the odd drive-by abuser because they turn up relatively infrequently, usually only when you’ve said something provocative or controversial.

But if you’re an outspoken public figure, the dynamic is completely different. It’s been said that “on a bad troll day” you can get 50 abusive messages an hour. That prevents you from using the Connect tab to connect with the sort of people you actually want to connect to.

Perhaps Twitter need to implement a variable setting which controls who you see or don’t see in the tab. The existing default will work for most people most of the time. A more restricted setting might limit this to your extended network, for example, those you’re following plus everyone they’re following.

If widely adopted, this might change the dynamic between Twitter celebrities and us normal people, limiting who can @message them, but maybe the existing dynamic is broken for at least some of the people, some of the time.

And the trolls will still troll, except their targets will no longer see them.

Another day, another twitter storm, this time directed at Caroline Criado-Perez for expressing an option a bunch of sexist trolls didn’t like. No matter how much you disagree with someone’s opinion, rape threats are never, ever acceptable, and it really shouldn’t need explaining why.

I’ve seen one or two people, notably James Desborough, defending the right to troll in the name of freedom of speech. But that ignores the fact that the trolls’ threats are themselves a silencing tactic. If you cannot voice controversial opinions without getting flooded with more threatening posts than you can cope with, then you’re effectively being censored.

But it’s very wrong to claim that there’s any one simple solution. Moderation of any online space (including Twitter) is very, very hard to get right, and far too easy to get horribly wrong. Even successful moderation policies in smaller communities don’t scale to something the size of Twitter. Unfettered freedom of speech can result in a playground for sociopaths where bad speech drives out good, but who gets to decide what’s good and bad?

Twitter’s big problem is flash mobs, when grown adults start acting like playground bullies. Many different groups are guilty of this. There are a lot of parallels on Twitter between the behaviour of clumps of misogynist trogodytes. and what happens when social justice types grab the torches and pitchforks in self-righteous fury. It doesn’t make much difference to the victim whether the mob is motivated by self-righteousness or sexist douchbaggery.

We should be very, very cautious about trying to use a quick technical quick fix for what’s essentially a social problem. It’s next to impossible to construct an automated abuse handling or crowd-based karma system that isn’t going to be gamed by the trolls and used as a weapon against their victims. And human moderation will involve subjective judgement calls which would have to take context into account.

I’m not saying that the status quo is the least bad option, just that we need to treat proposed solutions with caution, and be very wary of unintended consequences or unspoken agendas. I don’t want to live in an online world where people can be harassed and intimidated online, but neither to I want to see a situation where politicians and high profile media figures can shut down any criticism of their policies and views. What do we gain if we cannot call out a bigot for being a bigot?