Twitter is testing a feature that gives users advice for maintaining civility before they join a potentially controversial conversation
Context & Ripple Effects
This test is the latest step in a conversation-health program Twitter has run since it announced experiments tweaking retweets and replies to promote healthier conversations in late 2019. In May 2021 it rolled out improved prompts discouraging harmful language after studies showed the nudges worked, so today's feature extends the same logic one moment earlier: instead of warning users after they type something hostile, it coaches them before they enter a thread likely to turn contentious.
The significance is where the intervention sits — at the point of joining, not posting or reading. It pairs with the exit side of the toolkit Twitter was building in parallel, including Unmentioning on the web, which lets users leave conversations and stop notifications entirely.
First-order effects
- Users about to reply in heated threads get a civility prompt before composing, meaning Twitter's moderation surface moves upstream of the tweet itself for the first time in this experiment series.
Second-order effects
- Combined with exit tools like Unmentioning, entry-level coaching changes the calculus of pile-ons: potential repliers are deterred before piling in, while targets gain a way out once a thread turns, reducing the harassment load Twitter's own DM-icon test was criticized for amplifying.
Third-order effects
- If prompts keep proving as effective as Twitter's earlier studies suggested, content governance shifts structurally away from centralized takedowns toward user-facing friction — platforms shaping behavior through timing and choice architecture rather than enforcement.
The trend: Platform moderation is migrating from post-hoc enforcement to behavioral design, with interventions placed both before a user joins a risky conversation and at their exit from one.