OpenAI Releases New Voice Models for Live Conversation

OpenAI has announced new voice models designed for more natural, live interaction with artificial intelligence.

Speaking and Listening at Once

According to the developers, the key feature of the new models is that the system is able to speak and listen simultaneously. Previously, voice assistants typically operated on a turn-taking basis: first the user would say something, and then the model would compose and voice its reply.

What It Means for Simultaneous Translation

OpenAI notes that such a capability is particularly significant for real-time simultaneous translation, where you need to take in speech in one language and reproduce it in another at the same time, without any noticeable pauses.

The company sees this update as a step toward more natural interaction between people and voice assistants, bringing dialogue with artificial intelligence closer to an ordinary human conversation.

Source: TechCrunch