Twitter bots can be racist and offensive and were even accused of artificially boosting mentions of US President-elect Donald Trump during his campaign, but some are also good at curbing online hate speech.
That's according to NYU PhD candidate Kevin Munger, who conducted a two-month study using fake Twitter accounts to chide people who posted racist comments on the microblogging platform.
What he found was that people backed off when one of his "bot" accounts called out their racist Tweets, but only if the bots appeared to be white males who had lots of followers.
"I find that subjects who were sanctioned by a high-status white male significantly reduced their use of a racist slur," Munger wrote.
To create "high-status" accounts (ones with more than 500 followers), he bought followers. To vary their race, he chose handles that consisted of first and last names that were identifiably male and white or black according to another academic study, as well as cartoon avatars with different skin colors.
"In order to maximize the amount of control I had over the treatment, I used cartoon avatars for the profile pictures," he wrote. "This practice does not detract from the verisimilitude of the bot — using cartoon avatars on Twitter is not uncommon."
Each bot replied to racist tweets with messages like "Hey man, just remember that there are real people who are hurt when you harass them with that kind of language." When a bot masquerading as a white male sent the reply, the recipients future posts containing the n-word dropped by roughly 27 percent compared to a control group.
While Munger's results are promising for those who've given up hope that online comments can be stripped of racism, they also indicate the troubling fact that men who tweet racist comments apparently only care about the opinions of other white men.
At the very least, the fact that there was so much racist fodder for Munger's study confirms that like other social media platforms, Twitter is still far from cleaning up its act.