OpenAI is rolling out its Advanced Voice Mode to a wider audience, offering a more natural way to interact with its AI model, ChatGPT. This updated version allows users to interrupt responses mid-sentence and can adjust its tone and content based on the emotion it detects in the user’s voice.
Originally teased in May 2024 with the introduction of GPT-4o, Advanced Voice Mode has been available only to a select group of users since July. Despite initial safety concerns that led OpenAI to temporarily withdraw the feature, it has now been refined and is set to reach more users.
The new voice mode offers enhanced flexibility and responsiveness, fixing limitations found in the current standard voice mode. In the updated version, users can interrupt responses using their voice and the AI can adapt its replies based on emotional cues from the user’s tone. The update also includes improvements in the pronunciation of non-English words and allows users to personalise interactions by remembering specific facts about them.
OpenAI has introduced five new voices—Arbor, Maple, Sol, Spruce, and Vale—crafted with input from professional voice actors around the world. These voices have been designed to provide users with an engaging and pleasant conversational experience, featuring warm, textured tones.
Who Can Access It?
Advanced Voice Mode is currently available to Plus users, who pay $20 per month, and Team users, who pay $30 per month. Access will gradually extend to Enterprise and Edu subscribers in the coming weeks, though no specific deadline has been given. All Plus users are expected to have access by the end of autumn 2024.
However, users in the EU, UK, Switzerland, and a few other countries will not yet be able to access the feature due to geographic restrictions. Free-tier users will also not have access to Advanced Voice Mode for the foreseeable future, although the standard mode remains available to all paying users.
As OpenAI continues to expand the reach of its voice features, the new Advanced Voice Mode promises to offer a more dynamic and emotionally intuitive way to interact with AI.
