OpenAI unveils conversational GPT-4o

ChatGPT maker OpenAI has rolled out a faster and cheaper version of its flagship artificial intelligence (AI) model, dubbed GPT-4o, that includes a voice assistant mode.

GPT-4o (“o” for “omni”) can reason across audio, vision, and text in real time, in a step towards much more natural human-computer interaction, Mira Murati, OpenAI’s chief technology officer, said.

OpenAI live streamed GPT-4o’s debut at an event on Monday.

The new chatbot “accepts as input any combination of text, audio, and image, and generates any combination of text, audio, and image outputs.”

People can also interrupt the new voice feature during a conversation, quite unlike any of its peers, and get responses almost instantly, the company said.

During the demonstration, OpenAI chief technology officer Mira Murati and researchers conversed with the new ChatGPT to highlight the tool’s capabilities.

The chatbot appeared to translate speech almost in an instant, and even sang part of a story upon request.

It could also detect emotions from voice or facial expressions and rapidly switch between emotional tones and dramatic voices to robotic tones, in addition to singing.

The new model will offer free users a host of features that were earlier restricted to paid subscribers of ChatGPT, including the ability to search online for answers.

Users can speak to the chatbot and get responses in various tones while also ordering it to remember information that the user can retrieve later.

“The original ChatGPT showed a hint of what was possible with language interfaces; this new thing feels viscerally different. It is fast, smart, fun, natural, and helpful,” OpenAI chief executive officer Sam Altman wrote in a blog post.

“Talking to a computer has never felt really natural for me; now it does. As we add (optional) personalization, access to your information, the ability to take actions on your behalf, and more, I can really see an exciting future where we are able to use computers to do much more than ever before,” he wrote.

OpenAI will roll out GPT-4o’s text and image capabilities to paying ChatGPT Plus and Team users on Monday, and extend these capabilities to enterprise users later.

The firm plans to make the latest version of the “voice mode” assistant accessible to ChatGPT Plus users in the next few weeks.

OpenAI also said it is opening up its GPT store that hosts user-developed customized chatbots to all. The repertoire of GPT tools was earlier available to paying customers only.

The Microsoft-owned company’s announcement about the rollout of GPT-4o came a day ahead of Google’s flagship I/O developer conference, during which the search giant is also likely to roll out updates in the AI landscape.

Google has been locked in battle for dominance in the generative AI space with OpenAI and Microsoft.