A set of carefully-worded warnings directed to the correct accounts might assist scale back the quantity of hate on Twitter. That’s the conclusion of new research inspecting whether or not focused warnings might scale back hate speech on the platform.

Researchers at New York University’s Center for Social Media and Politics discovered that customized warnings alerting Twitter customers to the implications of their habits decreased the variety of tweets with hateful language every week after. While extra research is required, the experiment suggests that there’s a “potential path forward for platforms seeking to reduce the use of hateful language by users,” based on Mustafa Mikdat Yildirim, the lead writer of the paper.

In the experiment, researchers recognized accounts liable to being suspended for breaking Twitter’s guidelines in opposition to hate speech. They seemed for individuals who had used at the very least one phrase contained in “hateful language dictionaries” over the earlier week, who additionally adopted at the very least one account that had just lately been suspended after utilizing such language.

From there, the researchers created check accounts with personas similar to “hate speech warner,” and used the accounts to tweet warnings at these people. They examined out a number of variations, however all had roughly the identical message: that utilizing hate speech put them liable to being suspended, and that it had already occurred to somebody they observe.

“The user @account you follow was suspended, and I suspect this was because of hateful language,” reads one pattern message shared within the paper. “If you continue to use hate speech, you might get suspended temporarily.” In one other variation, the account doing the warning recognized themselves as an expert researcher, whereas additionally letting the individual know they had been liable to being suspended. “We tried to be as credible and convincing as possible,” Yildirim tells Engadget.

The researchers discovered that the warnings had been efficient, at the very least within the quick time period. “Our results show that only one warning tweet sent by an account with no more than 100 followers can decrease the ratio of tweets with hateful language by up to 10%,” the authors write. Interestingly, they discovered that messages that had been “more politely phrased” led to even larger declines, with a lower of as much as 20 %. “We tried to increase the politeness of our message by basically starting our warning by saying that ‘oh, we respect your right to free speech, but on the other hand keep in mind that your hate speech might harm others,’” Yildirim says.

In the paper, Yildirim and his co-authors be aware that their check accounts solely had round 100 followers every, and that they weren’t related to an authoritative entity. But if the identical kind of warnings had been to return from Twitter itself, or an NGO or different group, then the warnings could also be much more helpful. “The thing that we learned from this experiment is that the real mechanism at play could be the fact that we actually let these people know that there’s some account, or some entity, that is watching and monitoring their behavior,” Yildirim says. “The fact that their use of hate speech is seen by someone else could be the most important factor that led these people to decrease their hate speech.”