Google is reportedly nearing the introduction of an AI agent capable of operating a web browser to aid in automating everyday tasks for users. According to a report by The Information, the company is developing a "computer-using agent" under the codename Project Jarvis, which might be showcased as early as December. Sources indicate that Jarvis can respond to user commands by capturing frequent screenshots of the computer screen and interpreting these before executing actions such as clicking on buttons or typing into text fields.
The development of Jarvis is said to focus solely on web browsers, particularly Google Chrome, to assist with routine activities like research, shopping, and booking flights. This initiative aligns with Google’s ongoing efforts to enhance the functionality of its Gemini AI, the next-generation model of which is anticipated to be launched in December, according to The Verge. Recently, Gemini Live, Google’s AI chatbot, has expanded to support dozens of new languages, and its integration has been extended to Google Meet, Photos, and other applications.
The announcement of Jarvis follows closely behind Anthropic’s introduction of a similar yet seemingly more comprehensive feature for its Claude AI. This feature enables Claude AI to utilize a wide array of standard tools and software programs designed for general usage, and it is currently available in a public beta version.