Amazon announced a new AI model, Nova Sonic, designed to handle voice commands and data.
Amazon experts noted that the new model is capable of speech recognition, analysis, and composing sentences and voice phrases, and can engage in logical and realistic conversations with users. This model can also be used with the Alexa voice assistant in many electronic devices.
"The new model can work with our recently launched Alexa+ smart assistant," said Rohit Prasad, Amazon's senior vice president of artificial intelligence. "Nova Sonic is capable of conducting realistic conversations with users, taking into account silence or interruptions, and can answer users' questions directly."
Prasad noted that the new model is less error-prone than other AI models designed to process speech and audio, and can recognize a user's voice amidst noise. He also noted that Amazon plans to launch AI models capable of handling different types of data, such as images and videos, in the future.
According to the Multilingual LibriSpeech benchmark for speech recognition across multiple languages and dialects, Nova Sonic recorded an error rate of just 4.2% when dealing with French, English, Italian, German, and Spanish.