Google Innovates: Gemini and Video Avatars Revolutionize AI

Google Introduces Third-Party Connections in its AI Mode

Recently, Google launched a new feature for its AI Mode in Google Search, called Connected Apps. Currently available in the United States, this innovation allows users to directly link third-party applications during interactions with AI Mode. Among the first integrated applications are Instacart, which enables users to add ingredients to a shopping cart generated by the AI, Canva, for obtaining template suggestions without leaving the screen, and YouTube Music, which offers the ability to save a playlist created during the conversation.

This feature relies on the Model Context Protocol, a technology already used to connect third-party services to the Gemini application since June. Google announced it is collaborating with a wide range of partners and plans to expand this connectivity to other applications in the coming weeks.

Transforming NotebookLM into Gemini Notebook with a Cloud Computer

Google has decided to rename its NotebookLM tool to Gemini Notebook. Initially launched in 2023 under the name Project Tailwind, this research and document synthesis tool retains its autonomy while integrating more deeply into the Gemini ecosystem. Synchronization with the Gemini application is already in place.

This name change comes with a significant technological advancement: each notebook is now equipped with a secure cloud computer. This infrastructure allows for executing code to perform more complex data analyses. Currently, this feature is accessible to Google AI Ultra subscribers as well as Workspace clients with AI Ultra or AI Expanded access. A rollout for Pro subscribers is planned in the coming weeks. Additionally, these notebooks are expected to be integrated directly into the AI Mode of Search.

Google Vids Enhances Features with Avatars and Generative AI

Google has also enriched Google Vids, its video creation tool integrated into Workspace, with two major new features. The first, Gemini Omni, allows users to generate and modify videos from a text prompt and reference images such as a photo or sketch. Users can then adjust the videos step-by-step in plain language, modifying the background, adjusting lighting, or adding effects without having to start over.

The second new feature concerns personal avatars. These allow users to create a digital avatar that replicates their appearance and voice from a selfie and an audio recording. Text typed can be spoken by the avatar, eliminating the need for video shooting.

These new features are reserved for Google AI Pro and Ultra subscribers, as well as Workspace clients. However, access to personal avatars is restricted to adults in certain geographic regions. Each generated video is marked with an invisible SynthID watermark to ensure authenticity and content security.