Twitter runs a test prompting users to revise ‘harmful’ replies
In its latest effort to deal with rampant harassment on its platform, Twitter will look into giving users a second chance before they tweet. In a new feature the company is testing, users who use “harmful” language will see a prompt suggesting that they self-edit before posting a reply.
The framing here is a bit disingenuous — harassment on Twitter certainly doesn’t just happen in the “heat of the moment” by otherwise well-meaning individuals — but anything that can reduce toxicity on the platform is probably better than what we’ve got now.
Last year at F8, Instagram rolled out a similar test for its users that would “nudge” them with a warning before they post a potentially offensive comment. In December, the company offered an update on its efforts. “Results have been promising, and we’ve found that these types of nudges can encourage people to reconsider their words when given a chance,” the company wrote in a blog post.
This kind of thing is particularly relevant right now, as companies conduct moderation across their massive platforms with relative skeleton crews. All of the major social networks have announced an increased reliance on AI detection as the pandemic keeps tech workers away from the office. In Facebook’s case, content moderators are among the employees they’d like to bring back first.
We’ve reached out to Twitter for more information about the kind of language that triggers the new test feature and if the company will also consider the prompt for regular tweets that aren’t replies. We’ll update the story if we receive additional info about what this experiment will look like.