ChatGPT gains support for voice commands and image commenting

ChatGPT now offers audio or video interaction. This Monday (25), OpenAI announced that its chatbot Can adapt speech to text.

The interaction process is very similar to traditional chat, but the user must speak instead of typing. Just tap the record button to say the question or interaction. ChatGPT will then convert it to text and send it to the template.

Use your voice to join the back-and-forth conversation with ChatGPT. Talk to him on the go, ask for a bedtime story or settle a dinner table argument.

Sound on?? pic.twitter.com/3tuWzX0wtS

— OpenAI (@OpenAI) September 25, 2023

The technology behind the tool is Whisper, an artificial intelligence (AI) model that can adapt spoken speech to text. Just like virtual assistants, ChatGPT will also be able to respond using a synthetic voice.

According to OpenAI, this is possible thanks to a new text-to-speech AI model. The availability of the loudly responsive model will be limited.

Image interpretation

The other new feature is image interaction. The function is very similar to Google Lens, which can recognize elements of a shape or photo to search the web.

ChatGPT can now see, hear and speak. Available in the next two weeks, Plus users will be able to make voice calls and add images to conversations (all platforms) with ChatGPT (iOS and Android). https://t.co/uNZjgbR5Bm pic.twitter.com/paG0hMshXb

— OpenAI (@OpenAI) September 25, 2023

When uploading a file, ChatGPT will try to understand what the user wants to talk about and respond accordingly.. It will also be possible to use the editing tool to specify what needs to be identified by the AI, speak or write a question to complete the interaction.

In a way, this feature goes beyond Google Lens, as it’s not a simple web search. The versatility of OpenAI’s AI enables a more natural interaction, as if you were having a conversation with a person.

For security reasons, ChatGPT will not be able to recognize people. OpenAI states that the measure is a way to protect the public’s privacy. Without this, everyone would send photos to the model and ask “who is that?” he might ask. for example to get an answer.

Initially both new It will only be available to ChatGPT subscribers. The rest of the users will receive this news “shortly,” according to OpenAI.