Chinese people will be fine since all Chinese URLs are either ASCII compliant or use Chinese characters, which can't be confused with any ASCII characters.
To my understanding the unicode standard encodes an ASCII transliteration of an Unicode symbol to itself, but what about typographical similances? Wouldn't that be a hard problem? Perhaps there are two unicode characters that look exactly the same (using a given typeface) but have different transliterations. Or vice versa - two totally different looking characters share transliterations and gave false alarms.