OpenAI’s chatbot continues to improve. ChatGPT has just gained a new functionality allowing it to read the answers to your questions.
If not becoming even smarter than it already is, ChatGPT will at least become more practical to use on a daily basis. In a tweet posted on March 4, 2024 on X, the company behind the famous chatbot detailed how a new accessibility feature works: reading aloud.
How to make ChatGPT talk?
Available on the Android and iOS application and slowly arriving on the web, this tool allows, as its name suggests, to “vocalize” the responses provided by ChatGPT. Using a synthetic voice, the conversational robot can now speak, like a Siri or a Google Assistant.
To use the functionality, nothing rocket science. On mobile, simply initiate a conversation with the AI, let the robot finish its response then do a long press on the newly generated text. In the context menu which then opens under your fingers, simply click on “Read Aloud» so that a small audio player launches at the top of the page and begins transcribing the robot’s response.
It is also possible to skip 10 seconds forward and backward on the audio track with the dedicated buttons, or even activate the automatic transcription of messages so that all responses in a discussion are automatically read by the machine .
In addition to ChatGPT Voice
On the web, the manipulation is almost the same, since you just have to click on the small speaker below a response to ask ChatGPT to speak. Unfortunately, the feature is still gradually rolling out to the web client, so don’t worry if you don’t see it right away.
Whether on the web or on mobile, ChatGPT can automatically detect the language of the text and adapt its vocal rendering accordingly. The robot is capable of speaking 37 languages in total, including French with a touch of American accent.
This is not the first time that ChatGPT has explored the curious world of synthetic voices since last November the company unveiled ChatGPT Voice which did more or less the same thing, allowing you to ask your questions orally and to see the robot respond to it using the same method. The functionality unveiled by OpenAI at the start of 2024 is complementary to this, simply allowing certain responses to be read orally while asking questions verbatim.
Source : OpenAI – X/Twitter
0