We are halfway through OpenAI's 12-day celebration, and the sixth day will feature the “Advanced Voice Mode with Video,” which was first unveiled in May.

Every weekday through December 20, the AI Lab is announcing at least one product, service, or feature. So far, a variety of announcements have been made.

Today's OpenAI presenters include voice and vision experts Jackie Shannon, Michelle Chin, and Rowan Zellers, plus Chief Product Officer Kevin Weil.

ChatGPT can see you in real time while you speak in Advanced Voice mode and seems as natural as having a video conversation with a human.

As if this wasn't enough, you can also hear Santa's voice throughout December - and it seems he has a British accent. My 4 year old loves talking to the “AI” and is going to be fascinated with Santa's voice.

Video and screen sharing is rolling out today on the ChatGPT mobile app for all Teams, Plus and Pro subscribers outside of Europe.

Santa will be available wherever Advanced Voice Mode is available.

The demo was able to show ChatGPT's improved video, speech, and text memory features. Even with voice-only description, the system was able to memorize the name of the person on camera.

Because Advanced Voice is native multimodal, conversations take on a more natural tone than other models. It includes screen sharing as well as video, so the app can be shown to troubleshoot problems.

This allows you to show any app on your phone by selecting “screen sharing”. You can also open a message and ask ChatGPT for advice on replying to the message. It can also identify which app is open.

In another demonstration, Zellers set up a Pourover Coffee device and opened ChatGPT vision ChatGPT vision could identify the Santa hat and dripper he is wearing ChatGPT vision could identify the Santa hat and the dripper, and he walked through the steps of pouring the coffee.

ChatGPT's Advanced Voice maintained a natural, friendly voice throughout the demonstration, even changing tone and laughing as if it were human.

Advanced Voice with Vision is similar to Google's Project Astra, which Google updated yesterday with the announcement of Gemini 2.0.