Inspired by this previous question on UX SE: Good reasons to use bad UI? I asked myself the question, "Are there bad reasons to use good UI?" And just to clarify, I am talking about UI design and not UX design here.
The person who asked the first question gives the QWERTY paradox as an example, which I think only makes the argument from the existing user's perspective, which I believe that the convention wisdom is to weigh up the existing dependencies on the QWERTY standard and the continued cost of inefficiency for future users (which people don't look into enough for various reasons). This is basically the "if it ain't broke, don't fix it" argument.
It seems like the opposite statement of this question is also valid to ponder, as good/pretty UI sometimes takes away the problem of a bad underlying process, system or architecture temporarily, allowing users or companies to resist changing their solution even though a problem does exist.
So the question is whether a UI that reflects the underlying system/process is a better UI design compared to one that tries to cover the problems with a visual or interaction 'fix'? This is basically the "if there's a problem then you should fix it" argument. I believe that if you are only limited to working on the UI, then we should aim for the UI to reflect the way the system works so that the problems can be more clearly identified for a proper UX design to be done and reflect the way the user's mind works.