Google's Auto Browse in Chrome Shows AI Still Needs Human Hands
TL;DR Summary
A hands-on look at Google's Auto Browse for Chrome finds the AI agent can navigate and perform multi-step tasks, but it often doesn’t click when it should, misreads prompts (like seat selection), and requires user takeover for critical actions. While promising for automation, the tool is currently unreliable and raises security and trust concerns about relying on AI for purchases and browsing.
- I Let Google’s ‘Auto Browse’ AI Agent Take Over Chrome. It Didn’t Quite Click WIRED
- The new era of browsing: Putting Gemini to work in Chrome blog.google
- Google Chrome’s AI Agent Can Now Browse Web on Behalf of Users Bloomberg.com
- Google brings more Gemini AI features to Chrome browser CNBC
- Google Gives Chrome an AI Side Panel and Lets Gemini Browse for You CNET
Reading Insights
Total Reads
0
Unique Readers
15
Time Saved
11 min
vs 12 min read
Condensed
97%
2,293 → 62 words
Want the full story? Read the original article
Read on WIRED