OpenAI announced that the anticipated advanced voice mode feature on ChatGPT will be released at least a month later than expected due to safety and reliability concerns.
On Tuesday, the company revealed that the new voice capability of its bot would not be available to a group of ChatGPT Plus users later this month as originally planned, as it needs one more month to reach its “high safety and reliability” standards.
“For example, we’re improving the model’s ability to detect and refuse certain content,” the announcement says. “We’re also working on improving the user experience and preparing our infrastructure to scale to millions while maintaining real-time responses.”
OpenAI unveiled its new AI system GPT-4o in May, which includes several new features including an advanced version of its existing voice mode, making it capable of realistic voice conversation and of reasoning “across audio, vision, and text in real time”. This new model will have improvements compared to the existing voice mode that will allow users to have human-like natural spoken communications with computers, including a fast response time (an average of 320 milliseconds against GPT-4’s 5.4 seconds), and better perception of tones, background noises and emotions.
“GPT-4o (‘o’ for ‘omni’) is a step towards much more natural human-computer interaction,” the company said in a statement.
However, while OpenAI said the model should have an alpha release in July for a small group of ChatGPT Plus subscribers, the “exact timelines” for the next launching stages remain unsure. The company said it is planning “for all Plus users to have access in the fall”, although that will depend on the feedback received by the first group of users.
This is not the first obstacle OpenAI encounters with the launch of the advanced voice mode. Last month, actress Scarlett Johansson spoke out against the company claiming that its new speaking product imitated her voice without permission. The tech giant denied these allegations but still removed the voice.
Apple AI system in partnership with OpenAI potentially delayed too
Earlier this month, OpenAI and Apple announced they were partnering up to integrate ChatGPT technology in Apple devices, as part of Apple’s new AI system Apple Intelligence.
However, on Monday, after Apple was found in breach of EU competition rules under the new Digital Markets Act (DMA), the iPhone maker said that the release of its product with OpenAI could potentially encounter regulatory resistance.
