OpenAI has managed to stay in the news ever since the inception of the ChatGPT model and for revolutionizing Artificial Intelligence. The company has been constantly evolving the AI language model by bringing some advanced updates and features to better assist users with more relevant and contextual answers. To take the continuous improvement further, the company announced in May that it would soon be bringing in a Voice Mode on ChatGPT that would allow for more human-like conversations and personalization. The much-anticipated Advanced Voice Mode has now started rolling out to some users.
OpenAI is rolling out the new Advanced Voice Mode to limited paid ChatGPT users for testing
OpenAI showcased the development of a new Advanced Voice Mode for ChatGPT in May that would help users have more natural conversations in real-time, as well as eliminate delays or latency in interactions. It even landed itself in the middle of a controversy with its AI-generated voice called Sky, which sounded strikingly similar to Scarlet Johansson. However, the voice was taken down after the actress pursued the matter legally.
OpenAI announced on X that it had started rolling out the Voice Mode to selected ChatGPT Plus users for testing. It will keep on adding users and, hopefully, by this fall, allow all ChatGPT Plus subscribers to have access to the new feature. This AI tool is far more advanced than ChatGPT's current voice mode because it can understand emotions and varied situations and respond accordingly, making it feel more natural and human-like.
According to the post, OpenAI has undergone rigorous testing of the GPT-4o's voice capabilities with over 100 external red teamers across 45 different languages, making the version more quality stringent with greater privacy and security measures. The Advanced Mode will come with four preset voices: Juniper, Cove, Ember, and Breeze with ChatGPT Sky voice being dropped after the controversy.
The current preset voices do not resemble or impersonate any celebrities and would block any voices that do not match the four voices the Advanced Mode is programmed to speak in through a built-in system. To ensure no copyrighted content is generated or the voice is misused, OpenAI has placed filters and guardrails to prevent users from generating such content. Other multi-modal capabilities, such as video and screen-sharing, were demoed at the event but are not part of the current alpha and will be coming later.
The company plans to provide a more detailed report in August on the safety measures taken, the feature's potential, and its limitations. The selected ChatGPT Plus users will get an email containing the guidelines and a message in their mobile app, so if you are a paid ChatGPT subscriber, you better watch out for an email notification.
Follow Wccftech on Google to get more of our news coverage in your feeds.
