OpenAI has updated the voice feature of ChatGPT to enhance the naturalness of conversations in Korean. The update improves Korean intonation and voice expressions while adding two new voice options to enrich user experience.
On October 16, OpenAI announced that it has applied the voice model GPT-Live 1.1 to ChatGPT, enhancing intonation and voice expressions across multiple languages. This allows for more natural voice interactions in Korean.
New voice options for Korean users include a male voice named 'Hanul' and a female voice named 'Haetsal,' giving users the ability to choose their preferred voice. The voice feature can be utilized in various situations, such as organizing thoughts verbally, asking questions, or practicing foreign language conversations.
The conversational style during information searches has also been improved. Instead of repeatedly using phrases like 'checking,' the system now employs a variety of context-appropriate expressions, allowing for a more natural flow of conversation while waiting for responses.
ChatGPT's voice operates based on the GPT-Live model. It features a 'full duplex' structure, enabling users to ask follow-up questions while ChatGPT is responding. If a user pauses to gather their thoughts, the system waits patiently instead of rushing the conversation.
OpenAI introduced this real-time voice conversation method with the launch of GPT-Live in July. The latest GPT-Live 1.1 update builds on this foundation by adding new voices and improving multilingual intonation and conversational expressions during searches.
This update will be rolled out gradually to ChatGPT users worldwide, including those in Korea, and will be available to free users as well. Users can start a voice conversation by selecting the voice icon in ChatGPT.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.