In This Story
OpenAI is adding vision to ChatGPT’s advanced voice feature.
On the sixth day of its “12 Days of OpenAI” event, the artificial intelligence startup announced that ChatGPT can recognize objects shown through a smartphone camera or on the screen of a device, and respond with its Advanced Voice Mode feature. OpenAI teased the feature when it debuted its GPT-4o model in May.
Most ChatGPT Plus and Pro users and all Team users will be able to access the video and screensharing feature through the ChatGPT mobile app in the coming days, OpenAI said, while it expects ChatGPT Plus and Pro users in the European Union, Switzerland, Iceland, Norway, and Liechtenstein to get the feature soon. The chatbot’s Enterprise and Edu users will get video and screensharing in January.
Advanced Voice is powered by OpenAI’s natively multimodal 4o model, the startup said, meaning it can process audio input and respond in a natural and conversational way.
OpenAI also added a Santa voice as a preset voice in Advanced Voice Mode that is available to all mobile, web, and desktop app users. The feature, which can be accessed through a snowflake icon in ChatGPT, will be available worldwide until early January, the startup said.
At the beginning of the livestream announcement, the OpenAI team apologized for an hours-long outage that hit ChatGPT and its newly-launched video generator, Sora, the previous day.
The company released its text-to-video AI generator to some paid users earlier this week, which it called “critical to our AGI [artificial general intelligence] road map.” However, OpenAI chief executive Sam Altman wrote on X the following day that the startup had “significantly underestimated demand for sora,” and that it will “take awhile to get everyone access.”
As part of its holiday launch event, OpenAI has announced new products and features, including its o1 model out of preview and a $200-per-month subscription tier to ChatGPT. The chatbot is also now available to access through Apple’s Siri voice assistant.