Google unveils agentic AI and vibe-coded widgets for Android

At its “Android Show: I/O Edition” event on Tuesday, Google announced new Gemini Intelligence AI features for Android, including agentic task completion across apps, web browsing, form filling, speech dictation, and natural-language widget creation. The AI assistant can now handle multistep processes like copying a grocery list from notes and adding items to a shopping cart, with final confirmation required. The auto browse feature, first introduced experimentally in January for web-based tasks like booking appointments, is rolling out to Android. In late June, Gemini in Chrome will allow summarizing webpages and answering questions on Android. Form filling will leverage Personal Intelligence, an opt-in feature. Gemini is also coming to Gboard with Rambler, a multimodal dictation tool that transcribes speech and removes filler words. Users can build custom Android widgets by describing them in natural language, e.g., a meal planning widget. These features follow Material 3 design language and will first arrive on latest Samsung Galaxy and Google Pixel devices this summer, with broader availability later this year.