From Imagen 3 to DeepMind’s Mariner: Key features Gemini AI could unveil at Google I/O 2025

At Google I/O 2025, Google is reportedly planning to reveal significant updates for its Gemini AI platform, enhancing personalisation and productivity. Expected features include deeper Android integration and new collaborative tools, with the potential introduction of advanced functionalities.

Livemint
Updated30 Apr 2025, 08:15 PM IST
Google is gearing up to unveil a suite of new features for its Gemini AI platform at the upcoming Google I/O 2025 developer conference.
Google is gearing up to unveil a suite of new features for its Gemini AI platform at the upcoming Google I/O 2025 developer conference.

Google is gearing up to unveil a suite of new features for its Gemini AI platform at the upcoming Google I/O 2025 developer conference, according to the company’s latest newsletter sent to Gemini Advanced subscribers. The tech giant has hinted at transformative updates designed to enhance personalisation, boost productivity, and introduce new ways of engaging with the AI assistant.

As per a report by9to5Google, subscribers were told to expect “a wave of exciting updates” that promise a “more personalised assistant” experience. While Google has not confirmed specific tools or functions, speculation suggests that deeper integration with Android devices could be on the horizon. This might allow the AI assistant to handle a broader range of on-device tasks and interact more seamlessly with third-party applications.

In terms of productivity, Google is believed to be working on tighter integration between Gemini and its suite of workplace tools. Potential features could include collaborative environments akin to Google Canvas within Docs, or more advanced data visualisation capabilities in Sheets — possibly powered by Imagen 3, Google’s next-gen image generation model.

Also Read | Gemini AI’s next stop? Your smartwatch, car and tablet, reveals Sundar Pichai

The company also teased innovations in how users can interact with Gemini. Currently, interaction is limited to text, images, video, and voice. However, the real game-changer may come from the introduction of more advanced agentic features. Among the most anticipated is Project Mariner — an experimental AI agent developed by DeepMind that is designed to understand and operate across a wide array of web elements. Still in its alpha phase, Mariner could enable Gemini to perform complex, multi-step tasks in a user’s browser, marking a major leap in autonomous AI functionality.

Recent updates for Gemini subscribers have already included access to the Veo 2 video model and enhanced Gemini Live features.

With Google I/O 2025 just around the corner, developers and tech enthusiasts alike are eagerly awaiting what could be the most significant upgrade to Gemini since its launch.

Catch all the Technology News and Updates on Live Mint. Download The Mint News App to get Daily Market Updates & Live Business News.

Business NewsTechnologyNewsFrom Imagen 3 to DeepMind’s Mariner: Key features Gemini AI could unveil at Google I/O 2025
MoreLess