OpenAI delays release of new ChatGPT Voice Mode

It was arguably the centepiece of OpenAI’s big Spring Update event, at least alongside the new GPT-4o foundation model.

But the new Voice Mode the AI company showed off for ChatGPT — which allows for a more naturalistic, conversational, dare I say, “human” experience with the AI chatbot, complete with emotional inflection and the ability to handle interruptions — is officially being delayed by at least one month from its planned rollout date of late June to a “small group” of ChatGPT Plus users.

That means we shouldn’t expect Voice Mode until late July or early August at the earliest, and even then, it will only be available to a “small group of users.” The reason? OpenAI says it still needs more time to make sure the new Voice Mode can “detect and refuse certain content.”

As OpenAI posted on its X account today:

“We’re sharing an update on the advanced Voice Mode we demoed during our Spring Update, which we remain very excited about:

We had planned to start rolling this out in alpha to a small group of ChatGPT Plus users in late June, but need one more month to reach our bar to launch. For example, we’re improving the model’s ability to detect and refuse certain content. We’re also working on improving the user experience and preparing our infrastructure to scale to millions while maintaining real-time responses.

As part of our iterative deployment strategy, we’ll start the alpha with a small group of users to gather feedback and expand based on what we learn. We are planning for all Plus users to have access in the fall. Exact timelines depend on meeting our high safety and reliability bar. We are also working on rolling out the new video and screen sharing capabilities we demoed separately, and will keep you posted on that timeline.

ChatGPT’s advanced Voice Mode can understand and respond with emotions and non-verbal cues, moving us closer to real-time, natural conversations with AI. Our mission is to bring these new experiences to you thoughtfully.“

The news comes as a blow as OpenAI seeks to fend off the rise of rivals such as Anthropic, founded by former OpenAI team members. While Anthropic doesn’t have a voice output or input mode yet, it did just release a new foundation model, Claude 3.5 Sonnet, which appears to match or beat OpenAI’s GPT-4o model on various third party benchmarks.

OpenAI also facing off lawsuits and a growing tide of critics outside and within the company who are concerned about its safety measures (or lack thereof) surrounding its goal of artificial generalized intelligence (AGI), AI that outperforms humans at most economically valuable work. Another group of OpenAI employees attacked leadership for what they deemed to be overly restrictive separation agreements and equity restrictions, which OpenAI has since largely rescinded.

Other critics outside the company include actress Scarlett Johansson, who attacked OpenAI following the Spring Update event over a demo of an AI voice named “Sky” that she said sounded like her, after she refused to grant her likeness to the company. OpenAI and leaders have since rebutted her contention with receipts showing the company pursued the Sky voice actress separately from its overtures to Johansson. Nonetheless, the company disabled that AI voice.

But despite the growing tide of critics and setbacks, the company continues to impress and gain new adherents, such as the makers of new music videos and commercials with its unreleased video AI model Sora, healthcare startup Color which is integrating GPT-4o into a cancer screening app for clinicians, and signing lots of new enterprise accounts.