Google Unveils Live Video, Screen Sharing For Gemini AI At MWC 2025

Google showcased two new features for its Gemini Live AI application at Mobile World Congress 2025 in Barcelona.

The new features facilitate Gemini to engage in conversations, scan screens and environments, offer advice, and provide data in real time.

The first was the 'Live with Video' feature showcasing Gemini's conversational AI capabilities, where users could use their video camera to have the AI scan their surroundings and offer advice, suggestions, and information depending on the context of the questions asked by the user.

Google published a YouTube video showcasing the AI's capabilities, which depicted a user engaging in a conversation with Gemini about which glaze colour would look best on the vases they had fired in order to give it a 'mid-century modern' look.

The user showed Gemini a set of coloured tiles, which the AI was able to scan and provide a recommendation for. To further show the AI's sense of aesthetics, the user asked it about which of the colours it recommended would go best with the vases they had already painted, which it was also able to answer satisfactorily.

The second feature, 'Share Screen With Live,' showcased similar abilities from Gemini, but this time applied to screensharing instead of via a video camera. The user asked similar questions regarding style and aesthetics of the AI, this time asking Gemini what article of clothing they should pair with the jeans they were buying, displayed on the phone screen. The user asked further questions of the AI, asking it to adapt to new queries and build upon its initial answers, which it did so successfully.

The AI can seamlessly track the user's screen irrespective of scrolling through tabs, different apps, web pages, and settings.

The software tech company also held interactive demonstrations at MWC 2025 to show off these new features, developed by DeepMind, the company's AI division, which was a part of 'Project Astra.'