My Tech Reads

A General News & Tech Blog

Twitch will use machine learning to catch ban-dodging trolls

2 min read
Twitch will use machine learning to catch ban-dodging trolls

Twitch will use machine learning to catch ban-dodging trolls

Twitch introduces new machine learning features to help streamers protect their channels from people who try to avoid prohibitions. Dubbed “Suspicious User Detection,” This tool will automatically mark individuals suspected of being “possibly” or “may” prohibit Dodgers.

In the case involving the first, Twitch will prevent the message they send appear in the chat. It will also identify individuals for ribbons and whatever mods help them with their channels. At that time, they could decide whether they wanted to ban the person. By default, the possibility of repetitive trolls can send messages in chat, but they will also be marked by the system. In addition, Twitch said the creator had the choice to prevent them from sending any message in the first place.

“This tool is powered by a machine learning model that takes into account a number of signals – including, but is not limited to, behavior and user account characteristics – and compares that data on previous accounts is prohibited from the creator channel to assess the possibility of this account to avoid previous channel level prohibitions,” said A spokesman for twitching to Engadget when we asked about the signal used by the system to detect potential actors.

While twitching plans to turn a suspicious user detection for everyone, this tool will not automatically prohibit users for ribbons. It is with design because it is impossible to make machine learning tools that are 100 percent accurate in each context. “You are experts when they come to your community, and you have to make a last call about who can participate,” said the company in a blog post. “This tool will learn from the actions you take and the accuracy of the predictions must increase from time to time as a result.”

The introduction of tools follows summer where twitches struggle to contain phenomena called “Raids hate.” Attacks see evil individuals using thousands of bots to spam tract with hatred language. In many cases, they are targeting the creators of a marginalized community. Raids hate is a feature that often occurs on platforms that some creators walk away from twitches for a day as protest from the lack of company action.

Leave a Reply

Your email address will not be published. Required fields are marked *