OpenAI announced it is postponing the launch of its highly anticipated advanced Voice Mode for ChatGPT, a feature designed to enable sophisticated, real-time conversations with the AI. The company cited the need for further safety and reliability improvements, pushing the initial alpha release for Plus subscribers from its original late-June timeline to the fall.
The new Voice Mode, first demonstrated in May, represents a significant leap from the existing voice feature. It is engineered to understand and respond to users’ emotional cues, interpret tone, and handle interruptions, creating a more natural and fluid conversational experience. During the initial reveal, the technology showcased its ability to sing, tell jokes, and adapt its responses based on the user’s perceived emotional state. However, ensuring this capability is both safe and reliable at scale has proven to be a significant challenge.
In a statement, OpenAI explained that it is working to “improve the model’s ability to detect and refuse certain content” and to enhance the overall user experience. The delay also allows the company more time to prepare its infrastructure to handle millions of users accessing the feature simultaneously while maintaining real-time responses.
This cautious approach follows a period of intense scrutiny for the AI industry regarding product safety and ethical deployment. OpenAI itself faced controversy over one of its previous AI voices, codenamed “Sky,” which was criticized for its resemblance to Scarlett Johansson’s AI character in the film *Her*. Similarly, Microsoft recently walked back its “Recall” feature due to major security and privacy concerns. By delaying the Voice Mode, OpenAI appears to be signaling a greater emphasis on preemptive safety measures before releasing powerful new capabilities to the public. The company confirmed that other new features, including enhanced video and screen sharing functionalities, are also still under development.
English 
Deutsch