OpenAI announced that it is launching Advanced Voice Mode (AVM) to more of ChatGPT’s paying customers on Tuesday. This new audio feature makes ChatGPT sound more natural during conversations. At first, it will be available to customers on ChatGPT’s Plus and Teams plans, with Enterprise and Edu customers getting access next week.
With the launch, AVM is also getting a new look. Instead of the black animated dots shown in May, the feature will now be represented by a blue animated sphere.
Users will see a pop-up in the ChatGPT app, next to the voice icon, once AVM is available to them.
ChatGPT is also adding five new voices for users to try: Arbor, Maple, Sol, Spruce, and Vale. This brings the total number of voices to nine, close to Google’s Gemini Live. The other voices are Breeze, Juniper, Cove, and Ember. All the names are nature-inspired, possibly to emphasize that AVM is designed to make ChatGPT feel more natural to use.
Changes to OpenAI’s Voice Features and AVM Availability
A voice not included in this list is Sky, which OpenAI featured during its spring update. This voice led to a legal issue with Scarlett Johansson. The actress, known for playing an AI in the movie “Her,” felt Sky’s voice sounded too much like hers. OpenAI quickly removed Sky’s voice, explaining that the resemblance to Johansson’s voice was unintentional, even though some staff had mentioned the movie in tweets around that time.
One feature not included in this rollout is ChatGPT’s video and screen sharing, which OpenAI introduced four months ago during its spring update. This feature was meant to let GPT-4 process both visual and audio information at the same time. In the demo, an OpenAI staff member demonstrated how you could ask ChatGPT real-time questions about math written on paper or code on your computer screen. For now, OpenAI hasn’t provided a timeline for when these multimodal features will be available.
OpenAI has made some updates since the limited alpha test of AVM. The voice feature is said to be better at understanding accents, and conversations are reportedly smoother and faster. In our tests with AVM, we noticed some glitches, but the company says this has now been improved.
OpenAI is also adding some customization features to AVM for ChatGPT, including Custom Instructions, which lets users personalize how ChatGPT responds, and Memory, which enables ChatGPT to remember past conversations for future reference.
An OpenAI spokesperson mentioned that AVM is not yet available in several areas, such as the EU, the U.K., Switzerland, Iceland, Norway, and Liechtenstein.
Conclusion
OpenAI’s launch of Advanced Voice Mode (AVM) improves how ChatGPT talks with users by adding new voices and customization options. While some users can start using it, features like video sharing are still not ready. AVM aims to create a more natural experience, but it’s not available everywhere yet.
