It's not much consolation to me if I'm one of the 25% still being challenged.
The world really has more than enough heuristic fraud detection systems that most people aren't even aware exist, but make life miserable for those that somehow don't fit the typical user profile. In fact, the lower the false positive rate gets, the more painful these systems usually become for each false positive.
I'm so tired of it. Sure, use ML (or "AI") to classify good and evil users for initial triaging all day long, but make sure you have a non-painful fallback.
