AI Browser Automation Quiz

Answers and explanations appear as you take the quiz.

1. What core loop do AI browser agents follow at each step?
   • Observe the page, think about the next action, then act
   • Compile, link, run
   • Download, decompress, delete
   • Encrypt, sign, send
2. What behavior do Playwright locators provide before acting on a page element?
   • They guarantee the page’s content is safe to submit
   • They can auto-wait for actionability and re-resolve the target as the page changes
   • They infer the user’s intent without a goal
   • They prevent all layout changes from occurring
3. What does a vision-capable LLM typically receive as input on each step?
   • Only the page URL
   • A printed copy mailed to it
   • A screenshot and/or accessibility tree plus the goal and action history
   • The user's webcam feed
4. Which tool is commonly used to actually execute clicks and typing in the browser?
   • A spreadsheet formula
   • A PDF reader
   • A DNS server
   • Playwright or Chrome DevTools Protocol
5. What is a major safety concern with browser automation agents?
   • They act using your logged-in credentials and can take real actions
   • They use too little electricity
   • They only work offline
   • They cannot read text
6. Which is a sensible first workflow to automate with a browser agent?
   • Any long workflow, even if the target changes unpredictably
   • A bounded, repetitive task with clear checks and reversible steps
   • A payment flow with no review or confirmation
   • A task whose success cannot be checked
7. Why are role-based browser locators useful in automation?
   • They guarantee the page is accessible
   • They bypass authentication
   • They can target elements by accessible role and name rather than fragile screen coordinates
   • They prevent all layout changes
8. What should a browser agent do before a consequential action such as submitting a payment?
   • Click the first matching button automatically
   • Ignore page state
   • Treat model confidence as user permission
   • Confirm the target and action, then require appropriate authorization