It’s been a couple of days since OpenAI rolled out ChatGPT’s new advanced voice mode, and the small group of ChatGPT Plus subscribers given access to it seem pretty impressed so far. Various clips of the feature in action have appeared online, demonstrating its ability to sing, imitate accents, correct language pronunciation, and perform narrative storytelling.
An example of the latter can be seen in the below videos, in which X user @nickfloats asks ChatGPT to “tell me a story as if you’re an airline pilot telling it to passengers on a flight.” The chatbot jumps into action barely a second later, and even alters the audio to sound more like it’s coming from an intercom. ChatGPT struggled to accommodate more complex requests like layering on engine sounds, but the voice itself is clear and emotive and ChatGPT handles user interruptions well.
Guys im never talking to any of you ever again once gpt voice is released. I won’t need friends anymore. AI will tell me whatever I need to hear in any voice I want & it wont talk back or get mad when I interrupt it. Might even fuck around & fall in lovepic.twitter.com/GIRyhZYj9j
— Nick St. Pierre (@nickfloats) July 31, 2024
In a conversation uploaded to YouTube, ChatGPT says it can handle inputs in “dozens of languages,” but the exact number can vary “depending on how you count dialects and regional variations.” One clip demonstrates the chatbot’s ability to correct the pronunciation of French words, giving specific pointers on adjusting inflection. Another language demo shows ChatGPT speaking Turkish after following a detailed request to tell an emotive story. While some Turkish X users noted that the accent didn’t sound native, it was able to complete the story request and react appropriately by laughing and crying at certain points.
The bot does a passable job with regional US accents, with one video running through a variety of examples that include New York, Boston, Wisconsin, and a stereotypical “valley girl.” Other videos also show ChatGPT’s advanced voice feature singing in different styles, producing a blues-style take on “Happy Birthday” and, amusingly, trying to imitate what animals like frogs and cats would sound like singing the same tune.
ChatGPT Advanced Voice Mode attempting various US regional accents pic.twitter.com/UvDeQUNHLp
— Cristiano Giardina (@CrisGiardina) July 31, 2024
A few different male and female-sounding voices were present across these demonstrations, though these notably don’t include the Scarlett Johansson-like “Sky” voice that was pulled from the service in May.
As for anyone who feels left out of these fun demonstrations, OpenAI spokesperson Taya Christianson told The Verge that advanced voice mode will be available to all ChatGPT Plus subscribers (which costs $20 per month) sometime this fall.
