Sure, but definitions aside, are we fine with it? I'm fine with it. I think there is a perfectly reasonable distinction between being cautious and wary of individuals based on statistical evidence (including race/gender/age/dress/demeanor), and treating them differently once you start interacting with them.

I.e. Every individual I talk to has a fresh, blank slate with me. Stereotypes and "profiling" only really apply until you start getting a concrete instance of said group, and deal with them in getting "fresh" data that is unique to them.