The limited number of ChatGPT Plus users who have had access to OpenAI’s new advanced speech mode for a few days appear to be rather pleased with it thus far. Online videos of the feature in action have surfaced, showing off its abilities to conduct narrative storytelling, mimic accents, sing, and pronounce words correctly.
In the videos below, X user @nickfloats requests that ChatGPT “tell me a story as if you’re an airline pilot telling it to passengers on a flight.” This is an illustration of the latter. In less than a second, the chatbot begins to behave and even modifies the audio to make it sound more like it’s coming from an intercom. Although ChatGPT found it difficult to fulfil more complicated requests, such as adding engine noises, the voice itself is expressive and clear, and ChatGPT does a good job of managing user interruptions.
ChatGPT claims to be able to handle inputs in “dozens of languages” in a YouTube chat, while the precise number may vary “depending on how you count dialects and regional variations.” In one video, the chatbot is shown correcting French word pronunciation and providing precise guidance on how to change intonation. Another language demo demonstrates ChatGPT speaking Turkish in response to a thorough request for an emotionally charged tale. The accent didn’t seem natural, according to some Turkish X users, but it could finish the tale request and responded correctly by sobbing and smiling at the right times.
In one video, the bot runs through a range of instances, including New York, Boston, Wisconsin, and the archetypal “valley girl.” The bot does a respectable job with regional US accents. Other videos demonstrate ChatGPT’s sophisticated voice feature singing in many genres, including a blues-inspired rendition of “Happy Birthday” and a comical attempt to mimic the sounds of frogs and cats singing the same song.
Throughout these demos, a few distinct male and female voices could be heard; however, they noticeably lack the Scarlett Johansson-like “Sky” voice that was removed from the service in May.
For those who feel left out of these entertaining demos, ChatGPT Plus ($20 per month) users will have access to advanced speech mode at some point this autumn, according to OpenAI spokesperson Taya Christianson, who talked with The Verge.
For more fascinating tech news, be sure to check out TechExposed.
