OpenAI is continuously expanding the functionality of ChatGPT. While the tool was previously limited to text input, it is now possible to speak with ChatGPT. This means that “prompts” can be recited, and answers can be read directly instead of reading them from the screen.
Voice control in ChatGPT was announced at the end of September 2023 and has been rolled out to users since then (source: OpenAI). Initially, the function was distributed to companies and was only available with the paid version “ChatGPT Plus.” Since the end of November, “Chat GPT Voice” has been available to all users free of charge. So, you no longer need a subscription.
OpenAI announced on Twitter/X that the feature is now available to all interested parties:
ChatGPT with voice is now available to all free users. Download the app on your phone and tap the headphones icon to start a conversation.
— OpenAI (@OpenAI) November 21, 2023
Sound on 🔊 pic.twitter.com/c5sCFDAWU6
The voice output works similarly to Alexa and Siri. You say a command, which the artificial intelligence then processes. In contrast to the solutions from Apple, Amazon, and others, ChatGPT’s responses are of a higher quality and more detailed. The voice output is available in the ChatGPT apps for Android and iOS and must first be activated in the “Settings.” Make sure that you are using the latest version of the ChatGPT app.
Once the feature is activated, you can talk to ChatGPT:
- If you can access the function, you will find a headphone icon at the top of the chat history overview.
- Once tapped, you will see a short loading animation.
- As soon as you see a circle, you can activate voice input by pressing it.
- Smaller circles indicate that ChatGPT is listening. You can then speak your prompt.
- An animated speech bubble indicates that the command is being processed.
- ChatGPT then reads out the answer.
- If you tap on the corresponding function, you can also display the answer in text form as usual.
ChatGPT Voice: The chatbot can speak
By default, ChatGPT responds in English. However, you can also change the setting and then, for example, hear statements in German and other languages such as Polish, French, Spanish, or Japanese. A voice command cannot stop ongoing responses. So, you must listen to what ChatGPT has to say until the end. OpenAI relies on its own “Whisper” model for speech recognition. The company is said to have worked with five voice actors for the implementation. Answers can be delivered in five different voices.
Comments closed.