OpenAI has officially begun the rollout of its much-anticipated Advanced Voice Mode for ChatGPT Plus and Teams users, marking an exciting step toward creating a more human-like interaction with artificial intelligence. This feature promises to elevate the user experience by enabling real-time, fluid conversations, powered by GPT-4o, OpenAI’s latest model that integrates text, vision, and audio to deliver faster and more dynamic responses.
According to OpenAI, the Advanced Voice feature is being introduced to Plus and Team users within the ChatGPT app over the course of the week. The company cheekily noted in a recent tweet that the voice can now say “Sorry I’m late” in over 50 languages, a nod to the lengthy development period this project has undergone. However, one notable aspect remains absent: the much-discussed “Sky” voice, which stirred considerable attention due to its uncanny resemblance to actress Scarlett Johansson. Following concerns raised by her legal team, OpenAI decided to pause the Sky voice’s release, insisting that any similarities between Johansson’s unique voice and Sky were purely coincidental.
In lieu of the Sky voice, OpenAI introduced five new voices: Arbor, Maple, Sol, Spruce, and Vale. These voices will be available in both Standard and Advanced Voice Modes, joining the previously released options of Breeze, Juniper, Cove, and Ember. Interestingly, the names seem to have been inspired by soap fragrances, which adds a quirky twist to the technological advancement. The new voices are designed to facilitate more natural conversations, incorporating emotional responsiveness and the ability to interrupt and switch topics seamlessly.
Another noteworthy enhancement is the incorporation of custom instructions and “memories.” This feature allows users to personalise their ChatGPT experience further, tailoring interactions based on individual preferences. Just as the text-based chatbot learns from user-provided instructions—such as names, occupations, and preferred response styles—the new voices will adapt to conversational patterns over time. This personal touch aims to create a more familiar and engaging interaction, enhancing the overall user experience.
However, users in regions such as the EU, UK, Switzerland, Iceland, Norway, and Liechtenstein will have to exercise patience, as the feature has yet to be rolled out in these areas. Enterprise and educational users can expect access to the Advanced Voice feature starting next week, as per OpenAI’s rollout timeline. The implementation will occur gradually, meaning not all users from supported regions will gain access immediately.
Additionally, OpenAI has refined the accents in popular foreign languages and enhanced both the conversational speed and smoothness of interactions. The design has also received a facelift, with a newly animated blue sphere visually representing voice interactions in real-time, replacing the previous minimalist black dot that signalled voice activity. This design choice not only improves aesthetic appeal but also offers a more engaging user interface.
As OpenAI fine-tunes its voice AI capabilities, the competitive landscape in this domain is intensifying. Google’s NotebookLM currently stands out with some of the most human-like AI voices, capable of simulating entire debates between AI-generated speakers with astonishing realism. Google’s tool can process up to one million data tokens, allowing users to upload a specific group of documents containing various types of information. From this data, NotebookLM can generate up to 10 minutes of audio featuring two AIs discussing the uploaded information, resulting in an extremely realistic interaction.
Moreover, Meta has also entered the market with its live assistant, Meta AI, though it remains in limited release. This assistant is designed for natural conversations and can process commands fluently, offering a voice that is noticeably more natural than the robotic tones typically associated with AI assistants. However, it still exhibits certain characteristics—like speech cadence and speed—that reveal its AI origins. Notably, reports suggest that Meta’s forthcoming chatbot will channel the personas of renowned actors like Judi Dench and Michael Cera, providing a different yet compelling option in the voice AI space.
As the rollout of OpenAI’s Advanced Voice Mode unfolds, users will likely embrace these advancements in conversational AI. While the absence of the Sky voice is notable, the introduction of new voices that aim for naturalness and emotional connection represents a significant leap forward in human-computer interaction.
OpenAI’s commitment to enhancing user experiences through personalised features and voice versatility signals a forward-thinking approach in a rapidly evolving technological landscape. As competitors like Google and Meta continue to push boundaries in voice AI, the ongoing developments at OpenAI will be closely watched by both users and industry experts alike.
