Twitter expands Safety Mode beta, enables you to auto-block toxic accounts

Twitter Safety Mode interface on a mobile device

As a public social media platform, Twitter receives backlash from time to time for not being able to promote healthy conversations. This has led to many of its users becoming victims of toxicity, harassment, and abuse.

To counter this problem, the company released Safety Mode back in September 2021. This feature automatically blocks toxic users, hides toxic replies, and cautions users before they jump into potentially heated conversations. Think of it like a shield against unhealthy conversations, rather than a tool to outright stop them. That said, Twitter did caution that due to the autonomous nature of the tool, there are some flaws too, with the most obvious one being the potential creation of echo chambers. Regardless, the company is now rolling out the capability as a beta in more countries.

As noted by TechCrunch, Safety Mode is being made available in the U.S., UK, Canada, Australia, Ireland, and New Zealand. Twitter expects more users to give their opinion about the feature, how well it works, and how can it can be further improved. Additionally, the company also says that it will alert users when it is the right time to use Safety Mode.

A key feature of Safety Mode is auto-blocking of toxic accounts. You can turn on Safety Mode for a specific period, with a maximum duration of seven days. Before that period expires, you'll see a summary of auto-blocked accounts and then decide whether to unblock them. As for which accounts to block, that is determined by Twitter algorithms which assess the relationship between the poster and replier based on several factors. For example, if the poster is in frequent contact with the replier, they won't be blocked.

Twitter claims that the results of its early testing with 750 people show that people want more help in proactively identifying when they are going to be attacked online. As such, it is now being made available to more countries with prompts as well. 50% of randomly selected users in supported markets will be able to leverage from it. It's currently unclear when, or if, the capability will be rolled out globally.