Google has teased the arrival of new Gemini features and capabilities will be unveiled at the upcoming Google I/O 2025 developer conference. According to a report, the information was shared with Gemini Advanced subscribers in the company’s April newsletter. The Mountain View-based tech giant teased a “more personalised assistant” and new productivity tools for users. Additionally, the company also hinted at newer ways of interacting with the artificial intelligence (AI) chatbot. Gemini subscribers were recently given access to the Veo 2 model and new Gemini Live features.
Gemini Could Receive New Capabilities at Google I/O 2025
9to5Google reports that the tech giant has informed Gemini Advanced subscribers that they can look forward to new features for the AI chatbot that will be unveiled at its annual developer-focused Google I/O conference.
“We’ll announce a wave of exciting updates that will allow you to experience a more personalised assistant, unlock enhanced productivity, and open up new possibilities for interacting with and leveraging Gemini,” the company said in the newsletter.
While the company did not disclose what these capabilities could mean, “a more personalised assistant” could refer to new features in the Gemini AI assistant in Android devices. It is possible that the AI-powered voice assistant will be able to connect to more apps, and complete more on-device tasks than it currently can.
Similarly, productivity features could refer to the integration of the chatbot with more Google apps and newer features in existing apps. For instance, the tech giant could add a Canvas-like collaboration mode in Google Docs or introduce the ability to generate visual representations of Google Sheets data beyond graphs and charts. With Imagen 3, the company is already in a position to offer the latter.
Perhaps the most interesting part of the teaser is the “possibilities for interacting with and leveraging Gemini.” Currently, users can interact with the AI chatbot via text, images, videos, and voice. However, agentic capabilities in Gemini are still limited to the Deep Research tool. At Google I/O, the company might unveil DeepMind’s Project Mariner.
Project Mariner is a Gemini-powered AI agent that can complete several tasks with the user’s browser. It is natively multimodal and can understand text, code, images, forms, and other web elements. It is currently a prototype in the alpha testing phase.
