Twitter has long been a haven for some of the worst people on the Internet. The high rates of harassment against women and people of color have been an issue for years. Russia abuses it to drive American opinion and wreak further havoc in our political sphere. And, more recently, Twitter has been a major harassment tool for white nationalists and Nazi trolls.
There’s been some progress on Twitter’s part; “alt-right” leaders lost their official verification status and many white nationalists later found their accounts suspended altogether. However, a new op-ed by Yair Rosenberg for the New York Times reveals just how much more work it has to do.
Rosenberg knows a lot of Nazi trolling. The Anti-Defamation League says he’s the second-most harassed Jewish journalist on Twitter during the 2016 election cycle. Since Twitter wasn’t doing anything about it, he decided to matters into his own hands in a pretty brilliant way.
And so last November, in the wake of Trump’s victory, I decided to turn the tables on them. My target? Impersonator trolls.
You probably haven’t heard of these trolls, but that is precisely why they are so pernicious. These bigots are not content to harass Jews and other minorities on Twitter; they seek to assume their identities and then defame them.
Rosenberg says these types of trolls are particularly effective. They create new accounts, take the avatar of the person, and insert themselves into Twitter convos to say bigoted, inflammatory remarks. To undermine these trolls and their effectiveness, he got help to create a bot to hit them with the truth.
Using a crowdsourced database of impersonator accounts, carefully curated by us to avoid any false positives, the bot patrolled Twitter and interjected whenever impostors tried to insinuate themselves into a discussion. Within days, our golem for the digital age had become a runaway success, garnering thousands of followers and numerous press write-ups. Most important, we received countless thank-yous from alerted would-be victims.
The impersonator trolls seethed. Some tried changing their user names to evade the bot (it didn’t work). Others simply reverted to their openly neo-Nazi personas. A few even tried to impersonate the bot, which was vastly preferable from our perspective and rather amusing.