ChatGPT’s new voice is so real, it might fool your mom
This new update brings a more fluid, expressive, and human-like speech to the AI assistant for paid users. According to the company, the assistant now uses subtler intonation, has a more realistic cadence of speech, and even emotional inflection, like sarcasm or empathy.
Also, there’s a new feature for ChatGPT Voice: real-time translation. With this update, you can ask ChatGPT to translate between languages on the go, and it will continue translating until you stop it yourself.
So basically, you can speak English to it while traveling in another country, and it would continue translating your words into that other country’s language. It will also listen to the other person who may talk with you and translate their words into English for you.
Image Credit – Solen Feyissa
This useful addition eliminates the need for switching different apps or fiddling with them. It allows for a more seamless conversation even when someone isn’t speaking your language. The Advanced Voice mode is available to paid ChatGPT users in all markets. To access it, you can just tap the microphone icon in the message composer.
However, OpenAI does underline the new voice mode isn’t entirely without, well, AI faults. You may experience some dips in audio quality or tonal consistency with some voices, including some odd changes in pitch (depending on what you’re talking about with the assistant, I can imagine this sounding a bit creepy, don’t you think?)
Even funnier (or creepier) you may hear some fake ads or random background music during the conversation. OpenAI said it’s working on fixing those with a future update.
But all in all, the update seems useful and should get rid of the feeling you’re talking with a strange robotic creature. Personally, I like the idea of refining AI to sound less robotic, but at the same time, I still believe we shouldn’t forget it’s not really a human we’re speaking with. Maybe the random background music can help with that.
