Should I test advanced javascript UX on acceptance tests?

I have a form that implements google maps autocomplete to fill all of the address inputs. The UX is like this:

  1. The user sees only one address input (which is the google maps autocomplete one).
  2. When the user selects an address, the input goes away, showing all of the address inputs separately (already filled) along with a map.
  3. The user can fill the inputs that weren’t filled with the autocomplete, change the position on the map, and click a button (Reload address) to open the autocomplete input again.

My question is, what to test on acceptance tests (using Cucumber)? I don’t think it’s a good idea to go low level and trigger javascript functions, I know I should only simulate what the end user would do. So, should I just act like there wasn’t this javascript stuff at all and just fill out the address form?