OpenAI introduces Advanced Voice Mode with 5 new voices: What is new

After rolling out the Voice Mode, OpenAI has rolled out an Advanced Voice Mode for paid ChatGPT users. ChatGPT’s enhanced audio capability, enabling more conversational interactions, will initially be available to Plus and Teams subscribers, with Enterprise and Education customers gaining access starting next week. The Advanced Voice Mode not only brings new voices, but also a new look to the ChatGPT voice mode and a few customisations.
ChatGPT expands its voice options with five new additions: Arbor, Maple, Sol, Spruce, and Vale, bringing the total to nine voices, including Breeze, Juniper, Cove, and Ember. The number of voices is nearing the Google Live voices. You might notice all these names are inspired by nature, which could be because the whole point of Advanced Voice Mode is to make using ChatGPT feel more natural. This feature enhances speech patterns, tone, and pitch, adding emphasis and prosody to create engaging and lifelike dialogues.
According to the company, ChatGPT’s voice feature is better at understanding accents now, and its conversations are smoother and faster as well.
The visual representation of OpenAI’s feature has also been updated, swapping the previously demonstrated black dots for a dynamic blue sphere.
OpenAI is enhancing Advanced Voice Mode with two ChatGPT customisation features: Custom Instructions, enabling personalised responses, and Memory, allowing ChatGPT to recall previous conversations for future reference.
The most controversial Sky voice, which was eerily similar to Hollywood actor Scarlett Johansson’s voice, is still missing. After the actor officially filed a lawsuit, the artificial intelligence company paused the Sky voice in May 2024. OpenAI swiftly removed the voice, dubbed ‘Sky’, after backlash, claiming it wasn’t intended to mimic Scarlett Johansson’s voice, despite staff tweets referencing the movie Her. For perspective, Her is a Hollywood movie, where Johansson has given voice to the AI assistant.
Johansson also talked about the importance of transparency and the protection of individual rights in the age of deepfakes and AI. She urged appropriate legislation to safeguard personal identities and ensure ethical practices in the use of AI technologies.
ChatGPT’s video and screen sharing feature, debuted four months ago, is also missing from the current rollout. That feature is supposed to let GPT-4o simultaneously process visual and audible information. OpenAI demonstrated ChatGPT’s potential to analyse visual and audio inputs concurrently, enabling real-time queries on handwritten math or on-screen code, though a release date for this multimodal functionality is currently unavailable.