OpenAI ups the game, loads up ChatGPT with voice, image capabilities

From helping solve a math problem to talking you through a recipe from the contents in your fridge, ChatGPT will do it all. (Representational Image)

OpenAI's immensely popular artificial intelligence (AI) chatbot can now see, hear and speak.

In a blog post, OpenAI outlined the latest changes to the ChatGPT which have given it a voice, so to speak, and the ability to see images.

You can show ChatGPT a picture and have a conversation with it about the image. You can do this while travelling or even solving math problems.

ChatGPT's voice feature is limited to iOS and Android, so far, while the image capabilities will be available on all platforms.

To enable voice interactions, you need to go into settings and then new features on the mobile app, where you will find an option for voice conversations. This is an opt-in feature since it will use your voice recordings to train the AI.

Voice is powered by OpenAI's text-to-speech model which can generate human-like audio from text and a few seconds of sample speech. OpenAI says they collaborated with professional voice actors to create unique voices.

Images allow the chatbot put an image in context. You can, for instance, use it to analyse graphs, see recipes or label the contents in your fridge.

Simply tap the photo button on the mobile app to capture an image or alternatively, browse the gallery and choose an image.

OpenAI says ChatGPT can analyse a wide variety of images from photographs, screenshots or documents.

ChatGPT Plus and Enterprise subscribers will get the new features in the next two weeks, while developers can get their hands on it soon after.