Google's AI-powered 'Auto Browse' feature for Chrome aims to revolutionize the web by automating browser tasks. The tool, currently available only to users who subscribe to Google's $20-a-month AI Pro and AI Ultra plans, uses an AI agent to open tabs and complete digital tasks with automated clicks. However, a recent hands-on test of the feature revealed that it still has significant room for improvement.
When testing Auto Browse, I was initially excited by its potential to simplify browsing but quickly became frustrated as the tool struggled to deliver accurate results. A prompt to book tickets to the SF symphony resulted in two seats being selected from separate rows, rather than a single pair of adjacent seats. Similarly, when asking Auto Browse to find a leather jacket on Depop, it picked three options without demonstrating any discernible understanding of personal style.
A camping trip planning task also ended in disappointment as the tool failed to efficiently search for suitable campsites within a three-hour drive from San Francisco. In each instance, I had to intervene and take over, using my own judgment to correct the mistakes made by Auto Browse.
The experience highlighted the limitations of relying solely on AI tools for browsing tasks. While they excel at technical aspects, such as searching and completing forms, they often lack the nuance and common sense that humans take for granted. The post-click internet, where humans are gradually ceded control to machines, risks losing the very essence of online exploration and discovery.
Google's efforts to harness AI for seamless browsing may ultimately lead to a more efficient experience, but for now, it seems that users will still need to balance technology with their own intuition and decision-making. As I concluded during my tests: "My meandering quests through the internet’s backwoods are certainly not the most efficient journeys... But they remain delightful to me, and I would never want to outsource that joy completely to a browser bot."
When testing Auto Browse, I was initially excited by its potential to simplify browsing but quickly became frustrated as the tool struggled to deliver accurate results. A prompt to book tickets to the SF symphony resulted in two seats being selected from separate rows, rather than a single pair of adjacent seats. Similarly, when asking Auto Browse to find a leather jacket on Depop, it picked three options without demonstrating any discernible understanding of personal style.
A camping trip planning task also ended in disappointment as the tool failed to efficiently search for suitable campsites within a three-hour drive from San Francisco. In each instance, I had to intervene and take over, using my own judgment to correct the mistakes made by Auto Browse.
The experience highlighted the limitations of relying solely on AI tools for browsing tasks. While they excel at technical aspects, such as searching and completing forms, they often lack the nuance and common sense that humans take for granted. The post-click internet, where humans are gradually ceded control to machines, risks losing the very essence of online exploration and discovery.
Google's efforts to harness AI for seamless browsing may ultimately lead to a more efficient experience, but for now, it seems that users will still need to balance technology with their own intuition and decision-making. As I concluded during my tests: "My meandering quests through the internet’s backwoods are certainly not the most efficient journeys... But they remain delightful to me, and I would never want to outsource that joy completely to a browser bot."