OpenAI, an artificial intelligence company, has initiated the deployment of its eagerly anticipated “Advanced Voice” feature to a limited number of ChatGPT users.
“Advanced Voice is being implemented for all Plus and Team users in the ChatGPT app throughout the week,” OpenAI announced in a post to X on September 24.
“It can also say ‘Sorry I am late’ in over 50 languages,” a reference to the voice feature’s delayed release, initially scheduled for earlier in the year.
Advanced Voice Mode is an enhancement to the most recent 4.0 iteration of ChatGPT. It facilitates more efficient and intuitive communication with the model and several improvements to the conversational style reminiscent of human interaction.
Arbor, Maple, SXol, Spruce, and Vale are five new voices that OpenAI has introduced as part of the new feature. These voices are in addition to the existing Breeze, Juniper, Cove, and Ember voice options.
Users of the ChatGPT Plus and Team tiers will be granted staggered access to the new voices, intended to enhance conversations’ human-likeness. This includes interrupting the conversation and swapping topics in the middle of a conversation.
OpenAI also incorporates “memories” and instructions upon releasing new voices. Users can implement personalized instructions that customize the chatbot to their preferences.
Additionally, the chatbot can learn from and “remembering” significant details from previous audio conversations.
However, it only sometimes functions as intended. The conversation experience is not yet optimized for use with an in-car Bluetooth or speakerphone, and OpenAI acknowledged in a FAQ that ChatGPT could be interrupted by sounds.
However, some users on X expressed dissatisfaction with the new voice options, stating that they were inferior to Altman’s controversial “Sky” voice model, which was ultimately scrapped following a contentious legal dispute with actress Scarlett Johansson.
Johansson stated that Altman approached her in 2023 to serve as the voice of ChatGPT; however, she declined the offer due to “personal reasons.” Upon Sky’s appointment as the voice of GTP 4.0, she expressed her “shock, anger, and disbelief” at the chatbot’s strikingly similar vocal characteristics.
Despite making a single-word post to X that read “Her,” Altman subsequently scrapped the voice and maintained that any resemblance was purely by chance. This post was a direct reference to the 2013 Spike Jonze film, in which Johansson voiced the operating system of an intelligent AI companion.
