Textual JSON data always uses the Unicode character set. In this respect, JSON data is simpler to use than XML data. This is an important part of the JSON Data Interchange Format (RFC 4627). For JSON data processed by Oracle Database, any needed character-set conversions are performed automatically.
Oracle Database uses UTF-8 internally when it processes JSON data (parsing, querying). If the data that is input to such processing, or the data that is output from it, must be in a different character set from UTF-8, then character-set conversion is carried out accordingly.
Character-set conversion can affect performance. And in some cases it can be lossy: Conversion of input data to UTF-8 is a lossless operation, but conversion to output can result in information loss in the case of characters that cannot be represented in the output character set.
If your textual JSON data is stored in the database as Unicode then no character-set conversion is needed. This is the case if the database character set is AL32UTF8 (Unicode UTF-8). Oracle recommends this if at all possible.
JSON data that is not stored textually, that is, as characters, never undergoes character-set conversion — there are no characters to convert. This means that JSON data stored using data type
BLOB suffers no character-set conversion.
Unicode.org for information about Unicode
Oracle Database Migration Assistant for Unicode Guide for information about using different character sets with the database
Oracle Database Globalization Support Guide for information about character-set conversion in the database