Google expands Gemini with Chrome screen‑selection and built‑in computer‑use tools

Google has rolled out several new capabilities for its Gemini AI assistant. In Chrome 149 a "Select from Screen" option lets users highlight any text or image on a web page and send it directly to Gemini, turning the browser into an interactive chat window. At the same time, Gemini 3.5 Flash now includes native computer‑use functionality, enabling AI agents to see, reason and act across browsers, mobile apps and desktop environments without a separate model. Google added safety layers such as optional user‑confirmation for sensitive actions and automatic halting of tasks that appear to contain prompt‑injection attempts.

The updates also extend to user‑facing features: Chrome’s new‑tab page now shows an AI Mode button that opens a Gemini chat for conversational answers, and the Gemini app introduced a "Daily Brief" that aggregates information from Gmail, Calendar and Tasks into a personalized morning summary for paid users. Together, these enhancements aim to make Gemini a more context‑aware assistant for both everyday browsing and complex, multi‑step automation tasks.