Google's 'Project Jarvis' Could Redefine AI for Web Browsing

Google is reportedly developing an AI system, codenamed "Project Jarvis," designed to manage tasks directly within a web browser. According to The Information, this "computer-using agent" could debut as early as December and can perform various tasks on users' behalf, such as conducting research, making purchases, and booking travel arrangements. The report cites three insiders who are knowledgeable about the project.

Powered by Google's future Gemini AI, Jarvis is specifically optimized for Chrome and aims to simplify web-based tasks. It accomplishes this by taking screenshots, interpreting them, and executing commands like clicking buttons or typing. Currently, the system requires a few seconds between each action, as noted by The Information.

Google isn't alone in developing such technology. Microsoft's Copilot Vision is also being designed to interact with webpages through natural conversation, and Apple is working on intelligence features for cross-app screen awareness. Meanwhile, Anthropic's Claude and OpenAI are also advancing similar capabilities, although some remain in the testing stages.

However, The Information highlights that Google's timeline for Jarvis is flexible. The company may release it to a limited group of testers first, allowing for improvements before a broader rollout.