OK, bear with me: You can use iSeriesPython for this.
Assuming you have a faithful byte-for-byte copy of the files in the
IFS, you can open them in Python in binary mode to get a stream of
bytes. Then decode as necessary. For example,
f = open('NotepadUnicode.txt', 'rb')
bytes = f.read()
unicode_data = bytes.decode('utf_16_le')
At that point, unicode_data will be a string of Unicode characters,
which you can then either process directly with Python or encode into
a more favorable encoding and save back to the IFS for something else
to process. For example,
output = unicode_data[1:].encode('utf-8')
f2 = open('UTF8.txt', 'wb')
(Note that the [1:] strips off the BOM.)
Seriously, if Python is on the table, give it a try!