Imagine your web browser suddenly gaining a mind of its own. It can book your flights, summarise reports, and even input data into your CRM while you're busy with other tasks. This isn't science fiction; it's Google's latest move to embed its Gemini AI directly into Chrome for the workplace. But as it promises to give you back your time, a chilling new feature suggests it might also be taking away your privacy.
Announced at Google Cloud Next, the "auto browse" capability aims to handle tedious web-based work. Need to compare vendor prices across ten tabs? The AI will do it. Preparing for an interview and need a candidate's portfolio summarised? Consider it done. Google insists a "human in the loop" must approve all actions, framing it as a tool to liberate you for more strategic thinking. Yet, a stark reality looms: studies show AI often intensifies work, as expectations for productivity skyrocket.
The Dark Side of Your New AI Assistant
While selling convenience, Google is simultaneously rolling out a powerful surveillance tool for IT departments. Dubbed "Shadow IT risk detection," this feature allows companies to monitor every AI tool and SaaS website employees use—sanctioned or not. It’s a security measure, Google says, designed to spot compromised extensions and "anomalous agent activity." In practice, it’s a corporate lockdown, preventing the organic, employee-driven adoption of tech that once revolutionised offices with tools like Google Docs itself.
"We are essentially leveraging corporate IT to shut down any other AI agents," the announcement subtly admits, highlighting a brutal power play. The feature grants IT teams a "Gemini Summary" of browser changes and even suggests new policies, effectively making the AI both worker and manager.
What Happens to Your Data?
In an era where Meta trains AI on employee keystrokes, Google is quick to offer assurances. It states that an organisation's prompts won't be used to train its AI models. The "auto browse" feature will first launch for U.S. Workspace users, activated by a policy toggle. Users can save common workflows as "Skills," accessible with a simple "/".
But the integration goes deeper. Google is expanding its partnership with Okta to prevent session hijacking and introducing Microsoft Information Protection (MIP) Integration. The goal is a sealed, secure, and controlled environment—where the only AI agent allowed is Google's.
The promise is a future of effortless productivity. The hidden cost may be a loss of autonomy, as your browser transforms from a window to the web into a monitored company terminal. The race for AI supremacy in the office isn't just about who works faster; it's about who controls the very tools of work.