ChatGPT can now talk. OpenAI on Monday released an update to its AI chatbot that can have spoken conversations and interact using images. Like Amazon’s Alexa, Apple’s Siri and other digital voice assistants, users can speak to ChatGPT and the bot will respond.
The voice feature “opens the doors to many creative and accessibility-focused applications,” OpenAI wrote in a blog post Monday introducing the new features.
ChatGPT’s new voice feature can be used to have conversations on the go, “request a bedtime story for your family or mediate a discussion at the dinner table,” OpenAI cites a few use cases.
OpenAI argues that ChatGPT’s synthetic voices are more natural than others used in popular digital voice assistants. There are five different options to choose from, including male and female voices. According to the report, the new voice feature is based on a new text-to-speech model that is capable of generating a human-like voice from text and a few seconds of speech samples. To create the voices, OpenAI says it worked with professional voice actors.
The technology behind it is also being used by Spotify for the pilot phase of its Voice Translation feature, according to OpenAI, to allow the platform’s podcasters to translate their content into different languages using their own voices.
Like other digital assistants, however, ChatGPT has problems with homonyms, according to the US daily New York Times. The paper asked the new ChatGPT how to spell “gym”; the answer was “J-I-M.” But one of the advantages of a chatbot like ChatGPT is that it can correct itself, the paper said. To the interjection, “No, the other kind of gym,” the bot replied, “Ah, now I see what you mean. The place where people work out is spelled G-Y-M.”
Users of ChatGPT will not only be able to converse with the chatbot in the future, however, but also, for example, take photos of things around them and ask the chatbot to troubleshoot why, for example, the grill won’t start. When presented with a photo, table or chart, ChatGPT can provide a detailed description of the image and answer questions about its contents. Or users can upload a photo of the inside of their refrigerator, for example, and the chatbot can suggest a list of dishes they can prepare with the ingredients on hand.
The success of Microsoft subsidiary OpenAI’s ChatGPT has created hype around AI. Fast-improving AI technology can summarize documents, write computer code, produce intelligible speech and even photos and videos by processing and synthesizing massive amounts of data. More and more companies are betting on the use of AI and trying to bring their own generative AI-based applications to market.
