tech
Gemini 3.5 Flash can now see and control your screen, and Google wants enterprises to trust it
Computer use is now a built-in tool in Gemini 3.5 Flash, replacing the standalone Gemini 2.5 computer use model with enterprise safeguards.
TL;DR
- Gemini 3.5 Flash now features built-in computer use, enabling AI agents to see, click, type, and scroll.
- This capability consolidates a previous two-model workflow into a single tool within the Gemini API and Gemini Enterprise Agent Platform.
- Developers can now activate computer use alongside other tools like code execution and search within Flash.
- Enterprise applications include automated software testing, multi-step browser tasks, data extraction, and internal tool navigation.
- New enterprise safeguards are available, including explicit user confirmation for sensitive actions and automatic halting of potential indirect prompt injection attempts.
- These safeguards are opt-in and Google recommends a 'defense-in-depth' approach.
- The integration aims to make AI automation more accessible and secure for enterprises, though the technology still faces limitations with unfamiliar interfaces and dynamic content.