Google is rolling out expanded Gemini features that let the AI assistant take actions directly on your device. The updates, shown at Google's pre-I/O Android event, integrate Gemini across Chrome, autofill, and native apps.
Google is making Gemini do more than chat. The company's latest push puts the AI assistant in position to actively control your Android phone, handling tasks you'd normally do manually.
The new capabilities place Gemini in multiple locations across the OS. Users will find the assistant in Chrome on Android, embedded in autofill suggestions, and available within individual apps. This broader distribution aims to make Gemini assistance more contextual and accessible without requiring users to open a dedicated app.
Google introduced Gemini Intelligence, a branding term for enhanced capabilities that use a distinctive visual treatment. The company didn't elaborate extensively on what separates Gemini Intelligence from existing features, but the label suggests Google is positioning these tools as a cohesive suite rather than individual functions.
The timing of these announcements at Google's pre-I/O showcase indicates the company is preparing for a broader reveal of AI capabilities ahead of its main developer conference. Android devices will see deepening integration of Gemini across system functions, moving beyond conversational AI toward practical phone control.
Google's strategy reflects industry momentum toward AI assistants that take direct action rather than simply provide information. By embedding Gemini throughout Android, Google is betting users will prefer having the assistant available in their current context—whether filling a form, searching in Chrome, or working within an app—rather than juggling between tasks and a dedicated AI interface.
The announcement suggests Google still has more to unveil. The company confirmed these are updates during a "pre-I/O" event, meaning the full scope of Gemini's expanded role on Android will likely become clear during the main conference.
Anthropic is deploying its most powerful AI model, Claude Mythos 5, to strengthen cybersecurity defenses. The model now powers Claude Security, a tool that scans codebases for vulnerabilities and suggests patches.
Khosla Ventures has invested in Discovery Loop, an AI startup founded by former Google leaders including Jeff Dean, designed to accelerate scientific experimentation through artificial intelligence.
Nvidia research demonstrates that AI agents can perform reliably and safely through fine-tuning techniques, even when the underlying model lacks inherent task proficiency.
Developers and tech professionals describe experiencing cognitive fatigue and decision paralysis from constant exposure to AI tools and discussions. The phenomenon raises questions about sustainable AI integration in knowledge work.