OpenAI has started rolling out its advanced Voice Mode feature for ChatGPT. This new feature offers subscribers a more natural and immersive experience with the AI chatbot. A select group of ChatGPT subscribers already have access to this feature. They can engage in real-time voice interactions with the AI.

All ChatGPT Plus subscribers are expected to gain access to this toolset by fall. OpenAI announced via the social media platform X that the advanced Voice Mode “enables more natural real-time conversations and allows users to interrupt ChatGPT at any time. It also ‘senses’ and responds to users’ emotions.”

Voice conversation support was introduced in ChatGPT last September. A more advanced version was publicly demonstrated in May. ChatGPT-4o uses a single multimodal model for voice capabilities. This replaces the three separate models used by the previous audio solution. As a result, there is less latency in conversations with the chatbot.

In May, OpenAI faced criticism over a voice that sounded like Hollywood actress Scarlett Johansson. Her career includes voicing the AI character Samantha in the film “Her.” The release of the advanced Voice Mode was delayed due to this “misunderstanding.” Although the company insisted the voice was not intended to imitate Johansson, it was still removed.