The new AI agent that Google is allegedly developing has the ability to take over your computer, but it’s not as horrible as it seems.
According to reports, this AI agent, dubbed Project Jarvis (ostensibly a nod to the JARVIS voice assistant from the Iron Man films), has been specially trained to take over a user’s web browser in order to perform duties such as conducting research, making purchases, or scheduling travel.
Project Jarvis will be a user-facing tool that runs in Google Chrome and automates routine web-based tasks, according to a report by The Information.
Google plans to unveil the AI agent and its upcoming flagship Gemini AI model, Gemini 2.0, as early as December.
Following the preview, Jarvis might be made available to early testers and could be used in Google’s marketing campaign that highlights the capabilities of its most recent Gemini model.
How will the AI in “Jarvis” function?
According to reports, Jarvis will operate by capturing “frequent screenshots” of the user’s screen and using the information to execute commands like pressing a button or entering text.
Laptops 1000Jarvis is described as operating “relatively slowly,” nevertheless, as it requires a few seconds to process information before acting.
Although Jarvis appears to be the futuristic technology we have been anticipating, the most recent study presents more questions than it provides answers.
For example, we do not yet know if Jarvis will run in the cloud or on-device, which raises serious privacy concerns.
Furthermore, it’s unclear if Jarvis will simply function with Google Chrome on the web or if it would also work with the Google Chrome app on mobile devices.
In an interview earlier this year, Google CEO Sundar Pichai expressed his appreciation for AI agents, saying, “I think about [agents] as intelligent systems that show reasoning, planning, and memory, possess the ability to plan ahead, collaborate with software and systems, and complete tasks on your behalf—and, most crucially, under your supervision”.