Asking people to be nice on Twitter makes people nicer on Twitter, the company has announced, hailing the success of an experiment that prompts users to reconsider tweets that might be hurtful or offensive.
Since 2020, the social network has tried to encourage users to be more considerate of others, by algorithmically spotting posts that appear aggressive or mean-spirited and prompting users to consider whether they want to send them, asking: "Want to review this before tweeting?"
Now, in a study co-authored with Matthew Katsaros at Yale Law School, Twitter researchers say the prompts do indeed work to encourage a change in behaviour - and that the change sticks around long after the prompt was shown.
"We found that out of every 100 tweets where users were prompted to reconsider, the following actions were taken: 69 were sent without revision; 9 were cancelled; [and] 22 were revised," wrote Twitter's Kathy Yang and Lauren Fratamico. Of those 22 revisions, only one was rewritten to be made more offensive, and more than a third were redrafted to be more palatable.
The effect lasts longer than simply the initial tweet; compared with users who didn't see a prompt, the incidence of repeat offences dropped by a fifth among those who were shown warnings. "This represents a broader and sustained change in user behaviour, and implies that receiving prompts may help users be more cognisant of avoiding potentially offensive content as they post future tweets," the researchers write.
The experiment's success was welcomed by social media experts. "It turns out that if you do even the bare minimum you can sizeably reduce how horrible Twitter users treat each other," said Ryan Broderick, an internet culture writer who runs the Garbage Day newsletter. "The fact Twitter, as a company, had to come up with a 'please don't cyberbully each other' warning for their users is a wildly embarrassing summation of how vicious they've allowed their website to become, but, hey, if it works, it works."
