Following an initial rollout in the United States, Google also expanded access to the Gemini Spark platform for AI Pro subscribers in more than 160 additional countries.
This integration marks a significant shift from remote web access to local browser control, enabling the AI to handle complex, multi-step digital errands.
According to Google, the system is designed to automate time-consuming tasks such as researching flight options to begin a booking or scheduling apartment viewings from saved listings.
To manage security risks associated with this level of access, the system requires manual user approval for sensitive actions like payments and includes built-in protections against prompt injection—a type of cyberattack where malicious instructions are used to manipulate an AI's behavior.
The expansion includes several mechanisms to help users manage the AI's activity, such as a dedicated indicator in the Chrome top bar that signals when auto-browsing is active.
These desktop capabilities follow recent updates that introduced local control for Mac applications and integrations with the Model Context Protocol (MCP), a standard for connecting AI models to data sources.
While the Chrome integration is currently limited to U.S.
users, the broader international release of Gemini Spark allows a global audience of professional subscribers to access the agent's growing suite of personal intelligence and code execution tools.