OpenAI rolls out lifelike voice mode to ChatGPT paid users

OpenAI has started rolling out Advanced Voice Mode for some paid ChatGPT users, giving them access to a GPT-4o-powered voice that resembles that of the AI in Her movie. The new Voice Mode offers a lifelike voice that feels like you are talking to a human. The Microsoft-backed company has released the alpha version to a handful of ChatGPT Plus users but confirmed a wider rollout to all paid users in the fall of 2024.

The new voice capabilities were first demonstrated at OpenAI’s event in May, shocking audiences with how hyper-realistic the voice sounded. The company also faced criticism because the voice was uncannily similar to Scarlett Johansson’s character ‘Sky’ in the ‘Her’ movie. However, it was nothing like the real-life voice assistants we have seen before. The chatbot could take questions on various topics and answer in a voice similar to a woman’s, while the demonstrators interrupted the bot to change the flow of answers.

According to OpenAI, GPT-4o is “a natively multimodal model capable of generating content and understanding commands in voice, text, or images.” It can understand complex problems more efficiently and faster than GPT-4, but its outstanding feature is the chatbot’s new voice, deemed a breakthrough in the AI world.

The Advanced Voice Mode was earlier set for a June release, but OpenAI delayed the rollout by one month to “reach our bar to launch.” Essentially, the company was addressing privacy concerns and working on the criticism of the chatbot for being too human-like. The company said it was “improving the model’s ability to detect and refuse certain content.” OpenAI has also “added new filters that will recognise and block certain requests to generate music or other copyrighted audio,” according to company spokesperson Taya Christianson.

If you are a ChatGPT Plus user, you will receive the new Advanced Voice Mode in the coming weeks — that is if you already do not have it.