When the "Set a Text Value" action is used with unicode characters directly, it works as expected:
Set Text [~text] to '“ ’ À ¢'
Write [Blue] '{TXT:~text}' to log
Will write "“ ’ À ¢" to the log.
However, if those same characters are read from a file using this action, they are not recognized properly:
Set Text [~text] to [C:\special characters.txt]
Write [Blue] '{TXT:~text}' to log
Will write "� � � �" to the log(ASCII code 65533).
I believe
this is the underlying issue.
Notepad, which I used to create the text file, uses ANSI encoding by default.
EDIT: Noted in the documentation for v1.7.3.2