Technology news and discussion on TalkFuse.
Gemini can now complete supervised tasks across Mac apps and files
Google has added computer-use capabilities to its Gemini desktop app for macOS, letting the AI carry out tasks across Mac apps, local files and websites. The feature makes Gemini more agent-like, as it can handle multi-step workflows while users continue using their screen and cursor.
The agent’s work appears in a live preview, and users can pause or stop it and approve actions before they happen. Google suggests uses such as filling in forms, organising batches of files and creating slideshows; the feature is initially available in English in the US to subscribers on AI Plus, Pro or Ultra plans. Google has not said when it will reach other markets or platforms.
In short
- Gemini can now perform tasks across apps and websites on a Mac.
- Users can preview, pause or stop its actions.
- Access is limited to US subscribers on eligible plans.
Steelman: both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
Supporters may see supervised computer use as a practical extension of an assistant: it can take on repetitive, multi-step work across apps and files, saving time while users retain visibility and the ability to intervene. A limited launch for subscribers also gives Google a way to learn how the feature works in practice before widening access.
The case against
Critics may argue that an AI acting across apps, files and websites raises meaningful privacy and security concerns, even with a live preview and approval controls, because mistakes or misunderstandings could affect sensitive information or make unintended changes. They may also question whether access restricted to paid subscribers and initially limited to US users will make the benefits unevenly available.
Explained simply
Gemini is Google’s AI assistant, and a desktop app is a version designed to run on a computer. Computer-use features let an AI interact with software in ways people normally do, such as opening files, using websites and moving information between apps.
This kind of assistant can carry out several steps in sequence, rather than only answering a question or generating text. That may help with routine tasks involving different apps, but it also raises questions about how much control people have over actions taken on their computer.
Google’s AI plans are paid subscriptions with different levels of access. The feature’s initial availability is limited by language, location, subscription and operating system, so it is not yet available to all Gemini users.
Read the full article at the source →
Originally published by TechCrunch as “Google gives Gemini the ability to use your Mac for you”.