OpenAI is always working on interesting new features and additions for its chatbot, and thanks to reverse engineering, we occasionally get a sneak peek at them before they launch. Over the weekend, a few engineers playing around with ChatGPT uncovered eight new voices within the AI chatbot that have yet to be made available to the general public.
In a series of posts on X, the ChatGPT users found a new “read aloud” icon in an experimental version of the web app. They eventually dug up eight new ChatGPT voices, each with their own distinct codenames, and you can hear all of them in the post below:
At the time of writing, there are four voices to choose from on ChatGPT: Breeze, Cove, Ember, and Juniper. There was a fifth voice, Sky, but it was removed due to how similar it sounded to a certain actress who starred in the movie Her.
As noted by TestingCatalog, the voices sound much more natural and expressive than the ones that the chatbot currently offers. They also do a better job of expressing non-verbal phrases and animal noises, as you can hear in the video above, and seem to understand how to emphasize words or phrases that are italicized or bolded by the user.
We have no idea when or if any of these voices will ever see the light of day. As TestingCatalog points out, there isn’t even room for eight more voices in the dropdown menu as it stands. Perhaps a major update is in the works that will include these voices.
