Initially rolling out in July 2021 and only on iOS, this new round of testing expansion will also include web and Android platforms (along with more users overall). As per Twitter’s original statement of intent, the feature is intended for replies to tweets as a way to suss out what types of replies to show more or less of. So it’s having users provide the data to influence the algorithm that will boost or hide replies, basically. The idea was met with a fair bit of uncertainty and pushback when it was first announced, with many users believing downvotes could be misused. For example, depending on the conversation and the overall audience, votes could be cast to boost or silence voices that align with users’ own biases. And while Twitter has stated that it’s “… learned a lot about the types of replies you don’t find relevant,” many are still concerned the feature will be abused. Or that it just straight-up won’t work as intended. In the recent announcement tweet, Twitter user Wordle Gummidge pointed out that “Every single time downvotes, obfuscated or not, have been trialed on platforms, they have led exclusively to increased silencing of marginalized voices,” and asked, “why do you think your implementation is going to be different?” Peasant Majicus was also unsure of Twitter’s ability to properly make use of the data provided by the test, asking, “Have you learned anything about the types of notifications that people don’t want to see?” going on to state that “… every time I click ‘See Less Often’ it does literally nothing.” Despite the expanded testing, upvoting and downvoting are still only available for select users on the web, iOS, and Android for now. Twitter has not given an estimate as to when it might be available for everyone yet.
