Google's Jarvis AI leaked accidentally ahead of its rumored December release

Google accidentally revealed its Jarvis AI, which is designed to take over a user’s computer to complete web-based tasks like making purchases, booking reservations, and research.

Initially reported as “Project Jarvis” in October, it was leaked on November 6 via an early version uploaded to the Chrome Web Store (via The Information). The tool is described as a “helpful companion that surfs the web for you,” using AI to perform such tasks, just like Anthropic’s Claude LLM.

Google, then, quickly removed the store page after. Although users are now unable to use the extension due to permissions issues, Jarvis is expected to launch alongside Google’s Gemini 2.0 AI in the coming month of December 2024.

The premise of Google’s Jarvis AI does sound exciting, especially now that we have voice-controlled personal assistants that can manage tasks across devices. We’ve seen a lot of new AIs, like ChatGPT’s advanced voice mode that sounds extremely lifelike and humanoid, so why not have one that could do your mundane tasks so you can focus on what matters more?

But, given Google’s not-so-good relationship with its users over privacy concerns, Jarvis AI may be the start of a domino effect that hinders Google’s advancement in the AI industry. Many would prefer an on-device AI or one with Apple-like privacy features.

Competitors like Anthropic have also launched their own AI agents, such as Claude, which includes features like the Claude 3.5 Sonnet-powered “computer use” that allows the AI to autonomously interact with your device. Or, in other words, it can browse or manage your files without constant supervision for mundane tasks.