Google is reportedly preparing to unveil a new artificial intelligence (AI) project, codenamed “Jarvis,” which aims to automate routine tasks within a web browser by the end of the year.
As outlined by The Information, Jarvis is described as a “computer-using agent” capable of completing tasks on users’ behalf, such as gathering research, making purchases, and booking travel. Powered by an updated version of Google’s Gemini AI, this tool operates within a web browser—primarily Google’s own Chrome—and is designed to streamline web-based activities by executing commands based on visual screen captures.
How Jarvis Works
According to sources familiar with the development, Jarvis functions by regularly capturing screenshots of the user’s screen, analyzing the content, and using this data to perform actions like clicking buttons or entering text. Although the tool requires “a few seconds” between actions in its current testing phase, Google plans to refine this capability.
Jarvis reportedly interprets each screenshot before taking the necessary action to accomplish tasks, allowing it to automate processes users would typically carry out manually within a browser.
This browser-based AI agent could preview as early as December, although The Information notes that this timeline is subject to change. Google may initially release Jarvis to a limited group of testers to identify and resolve potential issues before a broader launch. The project aligns with Google’s broader plans to enhance the functionalities of its Gemini AI model.
Earlier this month, Google’s Gemini Live AI chatbot introduced support for multiple languages and began integration across Google products, such as Google Meet and Photos, to assist with task automation and user interaction.
Tech Companies Push into AI Automation
Google’s entry into AI-driven browser automation follows similar advancements by other tech leaders. Microsoft, Apple, and Anthropic have recently unveiled or are developing their own AI models with capabilities akin to Jarvis.
Microsoft’s Copilot Vision will reportedly allow users to interact with webpages conversationally, while Apple is working on a system that can respond to screen content across applications. Anthropic recently released a public beta of an update to its Claude AI, which now includes computer-use skills but has been described as “cumbersome and error-prone.” Google’s Jarvis, though specific to Chrome and primarily focused on web-based automation, marks the company’s step into this competitive AI-assisted productivity space.
While Google’s Project Jarvis could make everyday online tasks much easier, but it also raises critical considerations around privacy, security, and data accuracy. As AI agents like this gain more control over personal and professional digital activities, companies like Google will need to prioritize safeguards that protect user autonomy and data integrity alongside these advancements.
Featured Image courtesy of Nathana Rebouças on Unsplash
Follow us for more tech news updates.