article --summary
The summary
Google has added computer use to the Gemini app on macOS, letting the assistant work across desktop apps and websites in a separate background window while the user keeps working. Users can watch a live preview, pause the agent, and approve actions when needed. The announcement landed just three days after Apple introduced stricter permission controls for macOS automation.
Apple's changes split Device Control and Data Access permissions into separate categories and require more explicit prompts for apps seeking broad system access. Users can approve individual requests rather than granting ongoing access, and they can revoke permissions while an agent is running. Developers building desktop agents will need to design around least-privilege access as a result.
Many details remain unannounced. Google has not explained how the background environment is isolated from macOS, how it keeps state across apps, which actions need approval, how it will operate under Apple's new restrictions, or what protections guard against prompt injection. It has also shared no performance figures or failure-handling details. The feature is rolling out in English in the United States to adult personal accounts on Apple Silicon Macs running macOS Sequoia 15.0 or later, with a paid Google AI Plus, Pro, or Ultra subscription required.
Editorial note: This is a summary of an article first published by The New Stack. The full piece is on their site.