Twitter Developers Attempting To Fix Gaps In Hateful Content Violations
Twitter developers have been working to combat hateful content on the platform, even more so since Elon Musk acquired the company and became the CEO.
“We’ve not only mitigated the recent surge in harmful behavior, but have reduced impression of this content in search by about 95%,” tweeted Yoel Roth, Head of Trust & Safety at Twitter.
Removing hateful content meant removing more than 1,500 accounts and reducing impressions on this content to nearly zero.
Roth explained in a tweet that impressions are Twitter’s primary success measurement for content moderation. This means how many times harmful content is seen by users. Sometimes it means dealing with a focused, short-term trolling campaign, such as the most recent case of hateful content.
Many people using the platform have reported hateful conduct in tweets and received notices saying it’s not a violation. Twitter is trying to fix that.
First-person reports, the hateful interaction happening to or targeting a person, and bystander reports, which happens to someone else, are treated different. Bystanders don’t always have full context of what happened. Many reports of Tweets that violate the bystander policy end up being marked as non-violative for lack of information.
Roth said Twitter is changing how these violations are enforced — not the policies, but the gaps in the policies.