Google open-sources AI that can distinguish between voices with 92 percent accuracy

Google researchers have open-sourced an AI system that can suss out speakers with state-of-the-art accuracy compared to previous methods.