Twitter Says that Its Prompts on Potentially Offensive Tweet Replies Reduced Negative Interactions by 30%

It is interesting to see what effect the slightest level of friction can have on negative online behaviors, and how it can reduce unnecessary anxiety or discord, just with a basic nudge.

Last February, Twitter relaunched its test of Warning prompts on tweet responses identified by automated Twitter systems may contain offensive remarks.

Twitter launched its first iteration of testing in May 2020, before deciding to suspend it during the US election period.

The new test, in an updated alert format, was released to select iOS users, and in a new experiment report, released this week, Twitter says that in 30% of cases where users were shown these prompts, they did in fact change or delete their responses, from In order to avoid the possibility of misinterpretation or abuse.

Which is a big sum – imagine if Twitter was able to remove 30% of negativity and abuse on its platform with a simple alert to each user.

This number also does not fully reflect the potential impact, as this means that in 30% of the cases viewed, the Tweet writer considered the recommendation, then re-evaluated their response. But in many of the other 70% examples, Twitter’s algorithms got it wrong, and may not have been offensive at all. This means that the benefit here, in terms of reducing anxiety, could be huge – and again, it’s amazing to think how such a simple claim can have such a big impact.

Although it shouldn’t be overly surprising. Back in 2020, Twitter too Add another popup alert Which appears when users try to reshare articles in their Tweets without actually opening the article link and reading it themselves first.

Prompt sharing article on Twitter

Three months after this implementation, Twitter reported that pPeople who saw these alerts were 40% more likely to open articles, while people who opened articles before retweeting, specifically, were 33% more.

Again, it’s just a small push, a small component of friction in the process. But it can have big consequences, as the speed of social media interaction often leads to less-than-thought-out responses, and once people are asked to re-evaluate, that is often enough to change user behaviors.

An extension of this could also be that more users are looking at their overall tweet responses more deeply, and what they actually share online. Which, on a large scale, could have a huge impact, and it will be interesting to see if Twitter eventually rolls out the alerts to all users.

Now, Twitter Extension of the initial testwith users in Brazil is now also preparing to obtain these claims.

But it can go further, and may end up being another important step in improving the in-app experience.

Leave a Comment