Why is unstructured data a "privacy minefield"?

Identifiers are hidden throughout free-form content, traditional column-based anonymization doesn't work, and fusing multiple sources enables re-identification.

Three compounding problems:

• Hidden identifiers: personal info is scattered through text, buried in metadata, or implied by context — extremely hard to detect and remove systematically.
• Traditional methods fail: techniques built for structured databases (e.g. "suppress this column") don't translate to free-form content where sensitive data appears anywhere, in any phrasing.
• Re-identification through fusion: even if each dataset is individually anonymized, combining several unstructured sources lets attackers correlate patterns and re-identify people.

Tip: In structured data you protect known fields; in unstructured data you must first find the sensitive bits — and you can never be sure you found them all.