Stuart,
You could spend a LOT of time considering how to support multiple character sets and how to detect which is being used in each case, and what fallbacks to try in cases where the default character set does not render as expected. Of course, many of these rendering errors are subtle and cannot be easily detected by a machine, so you need a way for a human to try out different character sets if one doesn't work, etc.... Pretty soon you've got different character sets used in different descriptions, and then all kinds of new problems, like needing to choose a char set before export, and developing a mapping and fallback table when converting... and suddenly your end users need to KNOW about character sets. In my opinion, it gets messy fast.
In 2.7 we have included the CSV validator, which includes a couple best-effort attempts to check for non-UTF-8 content. Turns out that this is not an easy thing to do - it's certainly not easy to detect in advance what char set is being used in a file if you don't know in advance.
If you really wanted to add something on the import side, it would be much easier to add an import option for an XSLX file (as opposed to a UTF-8 encoded CSV), where you can assume the character set used is windows-1252, the most renowned source of those infamously incompatible curvy "smart quotes" and such.
As for finding aid generation, it's all based on XML - which CAN work with other character encodings, but is certainly not guaranteed to do so. On the other hand,
per the W3C:
All XML processors MUST be able to read entities in both the UTF-8 and UTF-16 encodings
Meaning when it comes to reusing external libraries and tools, and thinking about ongoing maintenance and cross-compatibility over time, using a well-established encoding standard that is already THE de facto standard for web applications makes everything much easier.
In most cases: if you can convince your users NOT to use Microsoft products for cutting and pasting and preparing imports, I suspect that 90% of these types of issues will disappear, since most other companies / products made the switch to something unicode-compatible a long time ago.
That said, these are just my highly biased opinions, and please remember I am not a developer myself!