OpenAI has started rolling out the advanced voice mode for premium users of its AI chatbot ChatGPT. The audio feature is touted to be more natural and immersive and will be released through the week. 

The AI firm has said that it’s not yet available in EU countries including Iceland, Norway, Switzerland, Liechtenstein or the U.K. After ChatGPT Plus and Teams customers, the feature will be released for Enterprise and Edu users from next week. 

CEO Sam Altman announced the feature on X saying, “Hope you think it was worth the wait.” 

ChatGPT will also be getting five new voices that users can test called Arbor, Maple, Sol, Spruce and Vale, making the total number of voice options nine. 

While OpenAI had announced the advanced voice feature in May, the rollout was pushed back after actress Scarlett Johansson alleged that the demo voice called Sky was distinctly similar to her own in the 2013 film, “Her.” While the company responded to the claim saying they hadn’t modelled the voice after Johansson’s but paused using it after her legal counsel sent letters stating they didn’t have permission. 

For premium subscribers, they need to open the ChatGPT app to receive a notification once they have turned on access to the new feature. Then, the user can click on ‘Continue’ to go on. Once they start a new chat, a sound wave icon will appear next to the microphone icon and the ‘Message’ field. 

A small bump sound will be heard after a few seconds post which the circle in the middle of the screen that will then transform into a sky blue and white animation. The bot responds then. Users can choose a different accent or speed up a conversation.