A notable trend emerging in the realm of generative AI is the development of AI agents, and Google appears to be gearing up to introduce its own agent within the upcoming Gemini large language model (LLM).
This initiative, dubbed Project Jarvis, is designed to function as an AI agent integrated into the Google Chrome browser, allowing it to perform common tasks with a higher degree of autonomy based on simple queries or commands. Experts suggest that the addition of AI agents to Chrome could represent the most significant update to the browser since its inception in 2008, as reported by The Information.
Google has already illustrated the capabilities of Gemini as a shopping assistant and travel planner with its existing Gemini 1.5 Pro model, which powers the Gemini Advanced chatbot. In a future iteration, Project Jarvis is expected to handle an array of tasks such as navigating websites, filling out forms, and processing payments to fulfill user requests.
For example, while a Gemini 1.5 Pro model could respond to a query like “help me plan a vacation in December with a $2,000 budget,” Project Jarvis might take it a step further by saying, “plan me a vacation in December with a $2,000 budget. Book the flights and hotel, then email me the details,” as highlighted by Tom’s Guide.
The name “Jarvis,” which stands for “Just A Rather Very Intelligent System,” is inspired by Tony Stark’s AI assistant from the Iron Man series. Although the official name has yet to be confirmed, this codename is quite fitting given the enhancements being integrated into Google’s model.
Google first revealed information about AI agents during its Google I/O developers conference in May, emphasizing that this technology would aid in processing both speech and video content, while also improving response times in conversations.
It is speculated that Project Jarvis could be a component of Google’s next LLM, likely named Gemini 2.0. The announcement for this generative AI could occur in early December, although details about its accessibility remain unclear, according to The Information.
As Google moves toward unveiling its own AI agent, other companies have already introduced similar features. Notably, Anthropic recently showcased its Computer Use agent, designed to allow the Claude LLM to perform tasks independently on devices, currently in beta testing. OpenAI has also initiated its Swarm Framework, a collection of agents collaborating to complete tasks, though this technology is still in the experimental phase.