Google accidentally leaked a preview version of Jarvis, its new artificial intelligence tool, by briefly publishing it in the Chrome extension store. According to The Information, although some users managed to install it before it was removed shortly after its publication, they were unable to use it as they couldn’t bypass the access permissions. The incident occurred just two weeks after Google’s plans to launch this web assistant became known.
Subscribe to the Softonic newsletter and get the latest in tech, gaming, entertainment and deals right in your inbox.
Subscribe (it's FREE) ►With an advanced version of Gemini as a base, Jarvis is responsible for browsing the Internet and performing actions such as gathering information, purchasing a product, or booking a flight. This AI agent will be launched in December 2024, and could become a top assistant by simplifying online tasks and freeing up time for more complex activities.
The competition to develop assistants that can anticipate user needs has become the new focus of innovation in the tech industry. Anthropic, for example, recently introduced a new tool for its AI Claude that can control the user’s computer. This system allows the model to analyze screenshots to learn from the user’s activity and act accordingly, even being able to control the PC’s mouse itself.
Apple is also not lagging behind in the race for smart assistants, with a similar feature for Apple Intelligence called Onscreen Awareness, which will be capable of observing user activity and reporting it to the system, being able to perform similar tasks later on.
Microsoft has also experimented in this field with Recall, an AI feature that stores PC screenshots (including passwords) to “learn about the user” and offer more personalized assistance. However, shortly after its announcement, Microsoft delayed its launch due to several privacy and security issues detected by some experts, relegating it to members of the Windows Insider program and assuring that it would arrive “later.”
