Abstract:
Google announced on Thursday that its macOS version of Gemini desktop application has officially introduced "computer operation" capabilities, allowing artificial intelligence to perform tasks on behalf of users, including operating native Mac applications and browsing the web in the background. Users only need to give instructions, and Gemini can display the operation process in a real-time preview in the chat interface. At the same time, the user's screen and mouse will not be occupied, and the user can still continue to handle other work.

This function belongs to the "intelligent agent" type of AI service, which means that AI no longer just answers questions or generates content, but can take actions directly on the user's computer. Gemini can complete multi-step tasks across local files, applications and the Internet. For example, users can have it fill out a form, perform the same action repeatedly on a batch of files, sort files into different folders based on specific rules, or create a slideshow based on information from files and online data panels.
Like other AI agents that provide similar capabilities, users can pause, stop Gemini's operations, and confirm or approve actions before they are executed. Google said that this feature is especially suitable for handling repetitive and tedious workflows, helping users reduce manual operation time.
However, it is not too early for Google to enter the game this time. Major AI model developers such as OpenAI, Anthropic and Meta have already launched agent-capable functions in their desktop applications. Google just announced this week the wider launch of Gemini's "Tasks" feature, which allows AI to act more like an agent to assist users in logging into websites, completing online purchases, or booking services. In addition, Google has only recently begun letting early testers experience the upgraded "CC" AI agent it released in September, which is designed to help families stay organized through email, calendar, chat and task management.
The "Computer Operation" function is currently only available in English in the United States. Users need to install the Gemini for Mac application and subscribe to the Google AI Plus, Pro or Ultra service. Google hasn't said when the feature will be expanded to other markets or operating system platforms.
Comments