Twitter to test option to hide a tweet's replies in June, suspends 3x more abusive accounts reported within 24 hours than last year, relies less on user reports
with an emphasis on doing so more proactively (without relying on reports from people). http://twitter.com/... Vijaya Gadde / @vijaya : Healthy conversations on Twitter is our most important goal. We've been working to improve Twitter's ability to proactively detect and enforce our rules to help reduce the burden on people using Twitter. Here's an update on our progress and what's in the works: http://twitter.com/... @reuters : Twitter suspends 100k accounts for creating new ones after suspension https://reut.rs/2Gqc0Hp pic.twitter.com/7e8bpYgpmp @find_evil : Chad's idea & work to crowdsource hands-on-keyboard support needed to appropriately handle these abusive (putting it lightly) tweets towards @IlhanMN within Twitter's reporting framework is an instantiation of something I've wanted for years now: Defensive bounties. #infosec http://twitter.com/... Hamza Shaban / @hshaban : “Using Twitter's search feature, Loder combined terms such as “Rope,” “Bullet,” “Noose” and “Hanged” with Omar's two Twitter accounts, revealing dozens of inflammatory tweets.” http://www.washingtonpost.com/ ... Agent Smith / @anonycraig : Omar said that threats on her life have increased after Trump tweeted a video on April 12 Twitter would've typically taken down the threatening tweets once they were reported, but the company left them up to enable potential law enforcement collaboration http://www.buzzfeednews.com/ ... Ben Collins / @oneunderscore__ : New from me: Racist conspiracy theories created by white nationalists about Notre Dame filtered into mainstream spaces in less than 24 hours. Platforms don't have the systems in place to stop it, and it's stirring up hatred of minorities. http://www.nbcnews.com/... @standuprepublic : White supremacists are literally fanning the flames of Islamophobia after the accidental fire at Notre Dame Cathedral. They use tragedy, social media, and disinformation to spread their hateful ideology. http://www.nbcnews.com/... Ethan Zuckerman / @ethanz : It was incredible to watch this happen in real time. Folks were spinning narratives of churches burning across Europe within a few minutes of the fire being reported. There's very little news that someone isn't weaponizing to push forward an agenda these days. http://twitter.com/... HashTigre / @hash_tigre : Bwahhh...okay, so basically Facebook with a 140 char. limit and no edit option. “Starting in June, we'll be experimenting with ways to give people more control over their conversations by giving them an option to hide replies to their Tweets.” http://twitter.com/... Matt Navarra / @mattnavarra : Twitter to launch a ‘hide replies’ feature in June as an “experiment” http://techcrunch.com/...Image: @wongmjane http://twitter.com/...
Context & Ripple Effects
This June test sits at the end of a long moderation arc: Twitter first experimented with a quality filter for threats and abuse back in 2015, then pushed that filtering out to all users the following year. What changes now is direction — instead of only filtering what reaches timelines, Twitter wants authors to control what attaches to their tweets, and wants its own systems to catch violations before anyone reports them.
The suspension numbers are the other half of the story: tripling the rate at which reported accounts are actioned within 24 hours, plus mass-suspending 100,000 accounts created to evade prior bans, signals enforcement capacity catching up with policy. The reply-hiding experiment previews what later became formal reply-limiting tools.
First-order effects
- Tweet authors gain a direct lever over their own conversations for the first time — the ability to hide replies moves moderation from Twitter's queue into the author's hands.
- Abusive accounts face faster consequences: with three times more reports actioned within 24 hours and proactive detection reducing reliance on user reports, the reporting bottleneck that previously delayed enforcement shrinks.
Second-order effects
- Ban evasion becomes the pressure valve: the 100,000-account sweep against re-registered accounts shows enforcement success immediately shifting abuse toward account creation itself, forcing Twitter to police signup behavior, not just tweets.
- Rivals face raised expectations — if Twitter normalizes author-side reply controls and proactive detection, platforms like Facebook come under comparable pressure to ship conversation-shaping features rather than relying on after-the-fact takedowns.
Third-order effects
- If the pattern holds, moderation stops being a back-office function and becomes product surface: the line from this experiment runs through Twitter's later experiments tweaking retweets and replies for healthier conversation, toward reply permissions and context prompts as standard design elements.
- Structurally, platforms are absorbing the cost of trust themselves — shifting from user-reported enforcement to algorithmic detection plus user controls means the burden of a healthy conversation moves off the people using the service and onto the company's infrastructure and product decisions.
The trend: Platform moderation is shifting from reactive, report-driven takedowns toward proactive detection combined with user-facing conversation controls, making trust a designed product feature rather than a support function.