How Twitter Can Crack The Eggs
Twitter has an abuse problem. There is a noisy horde who somehow thinks of it as sport to issue anonymous abuse and threats at others, including threats of rape and murder. That among other things this somehow helps promote “ethics in game journalism” is a mystery best not contemplated if you value your sanity. Do not willingly delve into the Hashtag that Shal Not Be Named, lest the raging pack of Sea Lions charge.
Yet there are two relatively simple measures Twitter can take to greatly reduce this horde of vitriol, anger, and threats: enable users to automute accounts without phone verification and “crack the eggs” by deanonymizing abusive sock puppet accounts.
Automute
Standard Twitter accounts occur in two forms under the hood, those with a verified phone number and those without. Phone number verified accounts don’t necessarily tie an account to a person (after all, you can get a burner for a few bucks), but they do tie an account to an expenditure of time and effort. This is why on black market services, phone verified accounts for other services tend to be more expensive.
At the same time, Twitter wants to make it easy for people to sign up, and requiring phone verification does impose a barrier. Yet there can be a compromise: Mute Unverified.
Mute Unverified would be a new opt-in option for those who are tired of spam and abuse. This would auto-mute all mentions, replies, or other communication from accounts without phone number verification, unless either the tweeting account is followed or the tweet is retweeted by an unmuted phone verified account. If an account is muted due to mute unverified, the user should be notified when a mention is muted (but at most once a day).
This would significantly reduce both twitter spam and abuse. Spam accounts should become significantly more expensive if a significant number of users add mute unverified, while abusers would quickly find that creating an account to send one or two rape threats now becomes substantially more expensive: each abusive egg requires a new burner number.
Crack The Eggs
Most of the Twitter abuse comes from anonymous, throwaway accounts. Yet these abusers usually have real accounts: after all, why heap abuse on someone on Twitter if you don’t participate on Twitter? And how many abusers are scrupulous about their use of a VPN service only to send their abuse?
Twitter can crack the eggs using the same techniques the FBI used to catch General Petraeus, then head of the CIA: common login analysis. When an account is blocked for abusive threats, threats which, in person, would qualify as true threats, simply identify the other accounts which logged in from the same IP in roughly the same timewindow.
This has a potential for false positives, notably from cafes and other shared networks, but the shared networks should be visible to Twitter thanks to the diverse logins. Twitter could also maintain an internal confidence level. One time may be coincidence, but if two, three, or more abusive throwaways are associated with a legitimate account, Twitter should crack the egg, revealing to the abuse targets the actual abuser.
And its easy for Twitter to identify most of the threats that need this treatment: simply imagine that a Liam Neeson “hero” is calmly reading it to you. If it chills your blood, like so many of the tweets addressed at Anita Sarkeesian, its a true threat and worthy of account deanonymization.
These two measures, taken together, should produce a substantial impact on Twitter’s abuse problem: it would directly raise the cost of abusive accounts and actually add some risk for those who chose to create abusive sockpuppets.
Right now, Twitter is a cesspit. Yet its their cesspit, and if Twitter wants Twitter to continue to work, they need to take reasonable concrete steps to dissuade abusive accounts.
Acknowledgements
I believe Randi Harper first proposed auto-mute as a way of silencing abuse. If you want to understand the practical reality of how twitter abuse works, follow her work.
