Google unveils Neural Expressive redesign for Gemini AI at I/O 2026

At its annual I/O developer conference in Mountain View, California, Google announced a comprehensive visual and functional redesign of its Gemini AI model, dubbed Neural Expressive. The new design language introduces fluid animations, vibrant colors, new typography, and passive feedback throughout the app. Gemini vice president Josh Woodward presented the updates during a keynote, emphasizing that good design is about both aesthetics and usability. The redesign transforms the chat experience from plain text walls to rich responses incorporating images, video, timelines, and interactive diagrams. It also integrates Gemini Omni, Google's new AI video generation tool, allowing users to input combinations of images, audio, video, and text to create and edit videos. Additional features include regional accents and dialects for the chatbot, a light/dark mode with blue gradients, and updated app icons. Google also announced Project Genie for generating 3D environments from Street View, updates to the Provenance AI-detection service, Co-Scientist for research, Weather Next for forecasting, and molecular modeling tools. Google DeepMind CEO Demis Hassabis stated that these advances will help unlock AGI's potential, calling the current era the foothills of the singularity. The conference also highlighted Google's recent collaboration with British designer Ross Lovegrove to co-create a chair using generative AI, and its work with the UK government on an AI planning tool.