ChatGPT's Exciting Advanced Voice Feature Expected to Launch 'Next Week'

OpenAI recently announced on X (formerly Twitter) that its highly anticipated Advanced Voice feature for ChatGPT will begin rolling out “next week,” initially to a select group of ChatGPT-Plus subscribers. This alpha test aims to gather user feedback before expanding the feature based on their insights.

Advanced Voice transforms user interaction by allowing natural conversation without relying on text prompts, similar to chatting with another person. First introduced in May during the launch of GPT-4o at the company’s Spring Update event, this feature stands apart from typical digital assistants like Siri and Google Assistant. Unlike these systems, which often provide scripted responses, ChatGPT’s Advanced Voice delivers nearly instantaneous, human-like replies in various languages. The GPT-4o model boasts an average audio response time of just 320 milliseconds, comparable to human conversational speed. In the demo video, viewers can see how the model engages with multiple users, improvises discussions in both English and Portuguese, and exhibits human-like emotions, including laughter.

Details on how participants will be selected for the alpha trial remain unclear, though they will need to be $20/month ChatGPT Plus subscribers. Initially set for a June release, the alpha was delayed to enhance the system’s content moderation capabilities and strengthen its IT infrastructure to handle expected user demand. As announced in June, a full rollout of Advanced Voice is not expected until at least this fall, and its timing will depend on ensuring the feature meets high safety and reliability standards.

Integrating natural conversation capabilities into ChatGPT marks a significant leap forward. This advancement reduces the necessity of a context window, easing hardware requirements and broadening AI's potential applications—particularly for users with mobility or dexterity challenges. Additionally, by simplifying interactions, this feature paves the way for broader acceptance of AI technology among users who may be familiar with voice commands like “hey Siri” but find prompt engineering daunting.