On September 24, OpenAI began rolling out its much-awaited “Advanced Voice” feature for ChatGPT, enhancing user interactions with faster, more natural voice conversations. The new feature is gradually becoming available to Plus and Team users over the next week.
This update introduces five fresh voices—Arbor, Maple, SXol, Spruce, and Vale—designed to make conversations feel smoother and more intuitive. Users can now interrupt or change topics mid-conversation, providing a more dynamic experience. The feature also playfully addresses its delayed release by saying “Sorry I’m late” in over 50 languages.
The Advanced Voice feature arrives with the addition of “memories,” allowing the AI to recall key details from previous chats, offering users a more personalized experience. However, OpenAI cautioned that performance might be affected when using in-car Bluetooth or speakerphone systems, where background noise could disrupt interactions.
Despite the excitement, some users expressed frustration over the absence of the planned “Sky” voice. This voice, reportedly similar to actress Scarlett Johansson’s, was scrapped following concerns raised by the actress herself. OpenAI CEO Sam Altman maintained that the resemblance was unintentional, but the cancellation sparked a broader conversation on AI voice ethics.
OpenAI’s latest update marks a significant step toward more human-like AI interactions, even as debate continues around voice design and imitation in artificial intelligence.