Riot Games Valorant

Riot Games Will Gather ‘Valorant’ Voice Chats to Train AI Moderator

Riot Games will start collecting in-game voice conversations between North American Valorant players on July 13 to train an AI moderator that can curb disruptive and hostile behavior. The background monitoring will enable the AI to launch in beta this year. Riot is almost certainly working with Spectrum Labs, it frequent partner in AI content moderation, though the startup went unmentioned in the official announcement.

Valorant Voice

Gathering voice data is the latest step in Riot’s AI moderation plans, first announced in a privacy policy update last April acknowledging Riot’s right to record and use voice data transmitted over its communications channels. The data collected won’t be used for assessing problems reported by players, the recordings are specifically for training the AI language models so that the moderator can learn to identify and evaluate future reports across every Riot Games product. Players uncomfortable with their chat possibly getting recorded and analyzed can opt out only by disabling voice chat, but the company emphasized that it is being cautious about employing the AI prematurely.

“Voice evaluation during this period will not be used for disruptive behavior reports. That will only begin with the future beta. And we know that before we can even think of expanding this tool, we’ll have to be confident it’s effective, and if mistakes happen, we have systems in place to make sure we can correct any false positives (or negatives for that matter),” Riot Games explained in the announcement. “This is brand new tech and there will for sure be growing pains. But the promise of a safer and more inclusive environment for everyone who chooses to play is worth it.”


Conversational AI Content Moderation Startup Spectrum Labs Raises $32M