Twitter says it is rolling out improved prompts that discourage “harmful” language, first to iOS and then Android, after studies showed prompts were effective
A year ago, Twitter began testing a feature that would prompt users to pause and reconsider before they replied to a tweet using …
TechCrunchSarah Perez
Context & Ripple Effects
Twitter had already worked on the structure and filtering of conversations, including a redesigned reply layout and a DM filter for questionable language. The new prompts move that effort into the moment before a reply is sent.
Twitter’s iOS users, followed by Android users, will encounter revised prompts before posting replies that the company’s studies associated with less harmful language.
The company shifts part of its safety intervention from reviewing or filtering content after it is posted to asking authors to reconsider it before publication.
Second-order effects
Twitter’s later civility guidance and Safety Mode tests can operate as escalation paths around the same reply behavior: prompts aim to prevent problematic posts, while blocking tools address repeat harmful accounts.
Reply participants face a more actively managed posting flow, making conversation quality a product-design responsibility alongside the platform’s existing layout and message-filtering changes.
Third-order effects
If Twitter continues combining pre-post nudges with reporting and account-level restrictions, platform moderation becomes a layered product system rather than a single post-publication enforcement function.
The pattern points toward social platforms treating interface design as a durable lever for harm reduction, with stronger interventions reserved for behavior that persists after prompts.
The trend: Twitter is building a graduated safety model that starts with behavioral nudges at composition and escalates to reporting and account restrictions.
After testing and improving prompts that ask you to review a potentially harmful or offensive reply, we learned that this feature can help encourage more meaningful convos. We're now rolling out these prompts on iOS and soon Android. https://blog.twitter.com/... https://twitter.c…
So apparently they tested this shit on me. I didn't know it wasn't a real feature, and it just went away one day. Sometimes I took the cusses out. Most times though, it was “I said what i said”. https://www.nbcnews.com/...
“Want to review this before Tweeting?” the prompt asks in a sample provided by the San Francisco-based company. Twitter users will have three options in response: tweet as is, edit or delete. - so we are getting an edit button but it is BEFORE a posting https://www.nbcnews.com/..…
“34% of people revised their initial reply after seeing the prompt, or chose not to send the reply at all.” This is proof that content can drive behavior change. It's also proof that OMG YES people read content. https://techcrunch.com/...
Also, can we get a broad exception for large companies? Because I really don't need twitter to tell me I'm being “mean” to Amazon. 🙃 https://twitter.com/...
Twitter won't and hasn't released anything on making sure it doesn't depend on language models that unfairly punish the marginalized https://twitter.com/...
I'd like to see Twitter go one step further and, when it detects someone trying to reply to a tweet, tells them to mind their own business https://twitter.com/...