Be part of our day by day and weekly newsletters for the most recent updates and unique content material on industry-leading AI protection. Be taught Extra
OpenAI lastly added long-awaited video and display screen sharing to its superior voice mode, permitting customers to work together with the chatbot in numerous modalities.
Each capabilities are actually obtainable on iOS and Android cellular apps for ChatGPT Groups, Plus and Professional customers, and can be rolled out to ChatGPT Enterprise and Edu subscribers in January. Nonetheless, customers within the EU, Switzerland, Iceland, Norway and Liechtenstein gained’t be capable of entry superior voice mode.
OpenAI first teased the characteristic in Could, when the corporate unveiled GPT-4o and mentioned ChatGPT studying to “watch” a sport and clarify what’s taking place. Superior voice mode was rolled out to customers in September.
Customers can entry video by way of new buttons on the superior voice mode display screen to begin a video.
OpenAI’s video mode looks like a video name like Facetime, as a result of ChatGPT responds in actual time to what customers present within the video. It may possibly see what’s across the person, establish objects and even bear in mind individuals who introduce themselves. In an OpenAI demo as a part of the corporate’s “12 Days of Shipmas” occasion, ChatGPT used the video characteristic to assist brew espresso. ChatGPT noticed the espresso paraphernalia, instructed when to place in a filter and critiqued the consequence.
It’s also similar to Google’s lately introduced Challenge Astra, by which customers can open a video chat, and Gemini 2.0 will reply to questions on what it sees, like figuring out a sculpture present in a London road. In some ways, these options are extra superior variations of what AI gadgets just like the Humane Pin and the Rabbit r1 had been marketed to do: Have an AI voice assistant reply to questions on what it’s seeing in a video.
Sharing a display screen
The brand new screen-sharing characteristic brings ChatGPT out of the app and into the realm of the browser.
For display screen share, a three-dot menu permits customers to navigate out of the ChatGPT app. They will open apps on their telephones and ask ChatGPT questions on what it’s seeing. Within the demo, OpenAI researchers triggered display screen share, then opened the Messages app to ask ChatGPT for assist responding to a photograph despatched by way of textual content message.
Nonetheless, the screen-sharing characteristic on superior voice mode bears similarities to lately launched options from Microsoft and Google.
Final week, Microsoft launched a preview model of Copilot Imaginative and prescient, which lets Professional subscribers open a Copilot chat whereas looking a webpage. Copilot Imaginative and prescient can take a look at photographs on a retailer’s web site and even assist play the map guessing sport Geoguessr. Google’s Challenge Astra may learn browsers in the identical approach.
Each Google and OpenAI launched screen-sharing AI chat options on telephones to focus on the patron base who could also be utilizing ChatGPT or Gemini extra on the go. However a majority of these options might sign a approach for enterprises to collaborate extra with AI brokers, because the agent can see what an individual is taking a look at onscreen. It may be a precursor to fashions that use computer systems, like Anthropic’s Pc Use, the place the AI mannequin will not be solely taking a look at a display screen however is actively opening tabs and applications for the person.
Ho ho ho, ask Santa a query
In a bid for levity, OpenAI additionally rolled out “Santa Mode” in superior voice mode. The brand new preset voice sounds very like the jolly outdated man in a pink swimsuit.
In contrast to the brand new options restricted to particular customers, “Santa Mode” is now obtainable to customers with entry to superior voice mode on the cellular app, the online model of ChatGPT and the Home windows and MacOS apps till early January.
Chats with Santa, although, is not going to be saved in chat historical past and won’t have an effect on ChatGPT’s reminiscence.
Even OpenAI is feeling the Christmas spirit.
