When OpenAI first demonstrated an incredibly realistic, near-real-time “enhanced voice mode” for its AI-powered chatbot platform ChatGPT in May, the corporate announced that the feature could be available to paying ChatGPT users inside weeks.
Months later, OpenAI says it needs more time.
In a post on OpenAI's official Discord server, OpenAI explains that it had planned to make the improved voice mode available in alpha to a small group of ChatGPT Plus users in late June, but that ongoing issues have forced the rollout to be delayed until sometime in July.
“For example, we’re improving the model's ability to detect and reject certain content,” OpenAI writes. “We are also working to enhance the user experience and prepare our infrastructure to scale to hundreds of thousands of users while maintaining real-time responses. As a part of our iterative deployment strategy, we are going to launch the alpha version with a small group of users to gather feedback and expand based on our learnings.”
OpenAI says the improved voice mode may not roll out to all ChatGPT Plus customers until fall, depending on whether it passes certain internal security and reliability checks. However, the delay is not going to impact the rollout of the brand new video and screen sharing features, which were demoed individually during OpenAI's spring press event.
These skills include solving math problems using an image of the issue and explaining various settings menus on a tool. They are designed to work with ChatGPT on smartphones in addition to desktop clients, reminiscent of the app for macOS that was made available to all ChatGPT users today.
“ChatGPT's enhanced voice mode can understand and reply to emotions and nonverbal cues, bringing us closer to real-time conversations with AI,” OpenAI writes. “Our mission is to bring these latest experiences to you with thought.”
On stage on the launch event, OpenAI employees demonstrated how ChatGPT responds almost immediately to requests, reminiscent of solving a math problem on a bit of paper held in front of a researcher's smartphone camera.
OpenAI's enhanced voice mode caused quite a little bit of controversy since the default voice, “Sky,” resembled that of actress Scarlett Johansson. Johansson later released a press release saying she had hired legal counsel to research the voice and get accurate details about its development – and she or he had refused repeated requests from OpenAI to license her voice for ChatGPT.
While OpenAI denied using Johansson's voice without permission or impersonation, it later removed the offending voice.
