Google plans AI browser assistant "Jarvis" to automate web tasks
Key Points
- Google is developing an AI assistant called "Project Jarvis" that will be able to navigate the Chrome browser on its own and perform tasks such as researching, purchasing products, or booking flights.
- Jarvis regularly takes screenshots of the screen, analyzes them, and then takes appropriate actions. Unlike similar systems, Jarvis is initially limited to the Chrome browser and is primarily aimed at consumers.
- The agent is still relatively slow, and Google faces the challenge of convincing users to trust it with sensitive data. Jarvis is scheduled to launch in December along with the new Gemini language model, but plans are not yet finalized.
Google is developing an AI system called "Project Jarvis" that can independently control Chrome and perform common web tasks, according to a report from The Information.
The AI assistant will be able to handle everyday tasks like searching the web, making purchases, and booking flights without user intervention. According to three people familiar with the project, Project Jarvis works by taking regular screenshots of the browser window, analyzing them, and then performing actions like clicking or entering text, The Information reports.
While similar to a recent computer automation system from Anthropic, Google's approach differs in focusing solely on Chrome browser control rather than broader program access. The company is targeting average consumers rather than developers and office workers.
However, the system still faces some limitations. Sources say the AI needs several seconds to "think" before each action, making it relatively slow. Google also needs to address user concerns about sharing sensitive data like passwords and credit card information with the system.
Google plans to announce the system alongside its new Gemini language model this December, but this isn't set in stone.
AI firms shift focus as language models hit their limits
The December timeline aligns with earlier reporting from The Verge about Google's plans to launch its Gemini language model. However, that report indicated Gemini might not deliver major performance improvements over existing AI systems.
This may be why AI companies are increasingly focusing on practical applications, as they appear to be reaching a plateau in raw language model capabilities despite continued efficiency gains. In response, they're looking for new ways to make their AI systems more useful.
The "Jarvis" name surfaced earlier in discussions of Google's AI strategy. Former Google UX strategist Scott Jenson criticized the company in May, saying it aimed to create a Jarvis-like assistant to keep users within Google's ecosystem, motivated by fears that competitors might do it first.
AI News Without the Hype – Curated by Humans
As a THE DECODER subscriber, you get ad-free reading, our weekly AI newsletter, the exclusive "AI Radar" Frontier Report 6× per year, access to comments, and our complete archive.
Subscribe now