One of the most popular aids adopted by users to reduce the pain suffered from the use of passwords is browsers’ autocomplete feature. This feature, caching username and password after getting the user consent and using them later for automatic completion, is available in all modern browsers but communication with the user asking consent is implemented in different ways. In this paper, we report on user studies comparing active communication with a blocking dialog box and passive communication with a non-intrusive toolbar. We found that a dialog box misled users to save passwords in public computers. Conversely, no security problem was observed with passive communication. Our exploration provides empirical evidence for the risks of preferring active communication for password autocomplete and other similar interactions and sheds light on many other aspects of password autocomplete.