Twitch's new Suspicious User Detection tool aims to stop ban evasion

Twitch has introduced a new tool designed to detect people who create new accounts to circumvent channel bans. Called Suspicious User Detection, this new feature is intended to help combat harassment — a recurring problem on the streaming platform.

"Suspicious User Detection, powered by machine learning, is here to help you identify [returning malicious] users based on a number of account signals," Twitch wrote in a blog post on Tuesday. "By detecting and analyzing these signals, this tool will flag suspicious accounts as either 'likely' or 'possible' channel-ban evaders, so you can take action as needed."

Speaking to Mashable about these "account signals," a Twitch spokesperson offered little elaboration, saying only that they include the suspected user's behavior and account characteristics. Such behavior and characteristics are then compared against previously banned accounts to find potential ban evaders.

This Tweet is currently unavailable. It might be loading or has been removed.

How Twitch's Suspicious User Detection handles these accounts differs according to its assessment of how likely it is that they're guilty. 

By default, "likely" ban evaders' messages will be blocked from appearing in chat so other viewers can't see what they're spewing. In contrast, "possible" ban evaders' messages will still be shown, however the account will be flagged to the streamer and moderators, who can then choose to restrict it from the chat. Users can also choose to adjust these automatic responses, as well as manually add users or turn off Suspicious User Detection completely. 

Unfortunately, while viewers may be shielded, streamers and moderators will still have to see every potentially hateful message regardless of how much they fiddle with the settings. A Twitch spokesperson told Mashable this is so they have the option of removing restrictions from incorrectly identified accounts, and completely banning hateful ones.

"One thing to prepare for, particularly around launch, is that no machine learning will ever be 100% accurate, which means there is a possibility of false positives and false negatives," said Twitch's blog. "That's why Suspicious User Detection doesn’t automatically ban all possible or likely evaders… The tool will learn from the actions you take and the accuracy of its predictions should improve over time as a result."

Assessing every potentially hateful user that shows up in your Twitch stream can be emotionally and mentally draining, particularly if there's a high volume of them. But at least Twitch's new tool will help quickly identify previously banned users and hopefully keep them out.