Twitch is introducing a new machine learning feature to help streamers protect their channels from people trying to bypass bans. The tool is known as Suspicious User Detection, and it automatically flags people it suspects are “likely” or “possible” tape frauds.
In cases with the former, Twitch prevents messages you send from appearing in chat. It will also identify these people for streamers and any mods that help them with their channel. At this point, they can decide if they want to block this person. By default, possible repeat roles can send messages in the chat, but they are also flagged by the system. In addition, according to Twitch, YouTubers have the option of preventing messages from being sent in the first place.
“The tool is based on a machine learning model that takes a number of signals – including, but not limited to, the behavior and account characteristics of the user – and compares that data to accounts previously banned by a YouTuber channel to assess the likelihood that the account will bypass a previous ban at the channel level, “a Twitch spokesman told Engadget when we asked about the signals the system uses to detect potential perpetrators.
While Twitch plans to turn on suspicious user detection for everyone, the tool doesn’t automatically block users from streamers. This is by design as it is impossible to develop a machine learning tool that is 100 percent accurate in any context. “You are the expert when it comes to your community and you should make the final decision who can participate,” the company said in a blog post. “The tool learns from the actions you take and the accuracy of its predictions should improve over time.”
The tool’s rollout follows a summer when Twitch struggled to contain a phenomenon called “hate attacks”. In the attacks, malicious individuals used thousands of bots to spam channels with hateful language. In many cases, they targeted YouTubers from marginalized communities. Hate robberies became so common on the platform that some developers left Twitch for a day to protest the company’s lack of action.
All products recommended by Engadget are selected by our editorial team independently of our parent company. Some of our stories contain affiliate links. If you buy something through one of these links, we may earn an affiliate commission.