OpenAI is postponing advanced Voice Mode feature till later this year

OpenAI has announced it will delay the release of Voice Mode. The company says it needs more time to develop and build the infrastructure to make millions of real-time calls simultaneously.

Voice Mode is a real-time chat feature for users to chat with ChatGPT. It was first demonstrated during the Spring Update for GPT-4o last month. OpenAI showed the ability to interact with the LLM, mimicking a real human being, including emotions, non-verbal cues and body language.

OpenAI’s plan was to roll out Voice Mode to a small group of ChatGPT Plus users this month. The company however did not manage to meet the deadline.

Employees are still training ChatGPT to detect and refuse certain comments, for example when somebody asks ChatGPT how to rob a bank. The company is also working on improving the user experience and upscaling the infrastructure so ChatGPT can interact with millions in real-time chats.

In order to achieve that, OpenAI will start with a small group of ChatGPT Plus users to collect feedback. The AI tech company hopes it will finish this phase as soon as possible, so that all users will have access to Voice Mode in the fall.

“Exact timelines depend on meeting our high safety and reliability bar. Our mission is to bring these new experiences to you thoughtfully,” OpenAI says in a statement on X.

Some people do not hide their dissatisfaction, calling the delay the ‘biggest rug pull in history’. Another user is joking that SSI stands for ‘Shipping Someday Inthefuture’. In reality, SSI is short for Safe Superintelligence, which is the name of the company a former OpenAI employee founded earlier this month.

Finally, OpenAI says it’s also working on rolling out new video and screen sharing capabilities. When we can expect to see these features remains uncertain. The ChatGPT desktop app for macOS launched last Tuesday.