You could do a sysread of the first N bytes and see if any are above 0x7F.

But only if you're looking for english text. For virtually any other language, the range between 0xa0 and 0xff is valid, too. And there are sure some encodings where 0x7f-0xa0 is also valid.