Masking Versus Redaction: Why Hidden Document Data Can Still Leak

The article explains that masking replaces sensitive values with altered but similar-looking ones, such as a card number ending in the last four digits, while redaction aims to permanently remove the underlying content. It notes that drawing black boxes over text is not true redaction, since copying text from the file can expose the hidden words. Detection methods such as pattern matching, name lists, and trained models each have gaps that can miss sensitive data. The author argues that writing a new file without the values is safer for external sharing than masking on display. Finally, it warns that masking reduces rather than eliminates re-identification risk, since combinations of details can still point to an individual.