Google Brings Gemini Computer Use to Mac, Letting the AI Operate Apps and Files
Google

Google Brings Gemini Computer Use to Mac, Letting the AI Operate Apps and Files

Google has given its Gemini assistant the ability to take direct control of a Mac, letting the AI work through local files, open applications and browse the web to carry out tasks that users assign to it.

The capability, which Google calls computer use, runs through the Gemini app for macOS. It puts Google in the race to build agentic AI that acts on a person's behalf rather than simply answering questions, according to a report by Mashable.

What Gemini Can Now Do on a Mac

Once a user gives Gemini an instruction, the assistant can move between apps, search through files and pull information from the web to finish the job. The feature works on Mac hardware broadly, including MacBooks, the Mac Mini and the iMac.

In its announcement, Google showed several examples. In one, Gemini takes a child's drawing of a butterfly and turns it into a polygon-style 3D model in Blender, the open-source modeling program.

A Background Window Users Can Watch

Gemini does its work inside a separate, isolated window that runs in the background. A live preview appears in the Gemini app, so users can follow each action as the AI takes it.

Google stressed that the Mac stays fully usable while this happens. People can keep typing, browsing and doing other work while the agent handles its assigned task alongside them.

Guardrails for an AI With Deep Access

Letting an AI agent reach into local files and applications carries obvious risks, from accidental changes to exposure of sensitive data. Google has added several controls aimed at keeping users in charge.

Users choose which apps Gemini is allowed to touch. They can pause the agent if it starts doing something unwanted, and they can approve key decisions before Gemini acts on them. The live preview also serves as a safety check, giving people a chance to catch mistakes as they unfold.

Google Arrives After Its Rivals

The launch is a catch-up move. Meta, OpenAI and Anthropic already offer comparable agentic tools that can operate a computer on a user's behalf, which leaves Google following rather than leading in this area.

The Mac release also arrives alongside other agent features from Google. The same week, the company added Gemini Tasks, a capability that lets the chatbot complete tasks for people using it in a browser. It is separate from Google Tasks, the company's existing to-do list app.

Who Can Use It and What It Requires

Google says computer use is rolling out immediately to English-language users of the Gemini app. Access is limited to paying subscribers on the Google AI Plus, Pro or Ultra plans.

On the hardware side, the feature needs a Mac with Apple Silicon running macOS Sequoia 15.0 or later. Older Intel-based Macs are not supported.

What the Launch Signals for the AI Market

Bringing an agent to Apple's platform shows that Google wants Gemini to work wherever its customers do, even on a rival's hardware. For a company whose AI competes for paying subscribers, reaching Mac users is a practical way to widen its audience.

For businesses and developers, the release adds weight to the shift toward AI that carries out multistep work across real software. It also puts permission controls and human oversight at the center of how these tools are judged. Whether Google can close the gap with earlier entrants will likely depend on how reliable and trustworthy its agent proves in daily use.

Source: Mashable

Topics:

JeffkomStory Team

Editorial, JeffKom Story

The JeffKom Story newsroom covers startups, founders, funding and the technology shaping what’s next.

Keep reading

BrightAI Raises $97M Series A Led by Smith Point Capital to Catch Equipment Failures
Fundraising

BrightAI Raises $97M Series A Led by Smith Point Capital to Catch Equipment Failures

More in Google