Gemini Spark can now browse Chrome for you
By Steven Van ·
Google is rolling out Chrome auto-browse in the U.S. first and bringing Spark itself to Google AI Pro users in 160+ more countries.
Gemini Spark, Google's personal AI agent that works in the background under your direction, picked up a Chrome integration that lets it actually act inside your browser rather than just reason about tasks. Spark can now handle tedious web errands directly, the examples Google gives are scheduling apartment viewings or researching flights, using your logged-in accounts and saved passwords the way you would yourself.
Real browser access, with guardrails
Letting an agent use your logged-in sessions is exactly the kind of capability that raises the stakes on safety, and Google built the integration with that in mind. Spark includes protections against prompt injection attacks, a known risk when an agent reads and acts on content from the open web, and it requires explicit user approval for sensitive actions like payments. The agent stays in the loop with you on anything consequential rather than acting fully autonomously.
Starting in the U.S., expanding from there
Chrome auto-browse is rolling out first in the U.S., with more regions planned. It's a sensible rollout order for a feature this sensitive, giving Google room to monitor real-world behavior before extending the same capability, and the same risk surface, to more markets and languages.
Spark itself reaches far more subscribers
Alongside the Chrome feature, Google AI Pro subscribers in 160+ additional countries now have access to Gemini Spark at all, a major jump from its earlier limited rollout. Combined, the two updates mean far more people now have access to Spark, and the people who do get a materially more capable agent than the one that launched with just background task execution.
Sources: Google.