Google is reportedly working on a new AI agent that can take control of your computer, but it’s not as bad as it seems. This AI agent, named Project Jarvis (in an apparent reference to the JARVIS voice assistant from the Iron Man films), would be specifically trained to take over a person’s web browser to perform tasks such as collecting research, purchasing products or booking flights.

According to a report from The Information, Project Jarvis will be a consumer-facing feature that will work in Google Chrome and automate everyday web-based tasks. The report states that Google plans to preview the AI ​​agent as early as December, along with its next flagship Gemini AI model, known as Gemini 2.0. After the preview, Jarvis could be made available to early testers and could be part of Google’s marketing campaign showcasing the prowess of its latest product. Gemini model.

How will the “Jarvis” AI work?

Apparently, Jarvis would work by taking “frequent screenshots” of the user’s screen and interpreting those images to perform actions like clicking a button or typing in a text field. However, Jarvis reportedly works “relatively slowly”, as he requires a few seconds of thought before acting.

Although Jarvis seems like the kind of futuristic technology we’ve been waiting for, the latest report raises more questions than it answers. For example, we don’t yet know if Jarvis will run on the device or if it will run in the cloud (a major privacy issue). Additionally, it’s unclear whether Jarvis will only work with Google Chrome on the web or also with the Google Chrome app on mobile.

Google CEO Sundar Pichai previously showed his admiration for AI agents during an interaction earlier this year, saying: “I view (agents) as intelligent systems that demonstrate reasoning, planning and memory. Are able to anticipate multiple steps, work on software and systems, all to accomplish something on your behalf and, more importantly, under your supervision.