Overview
Auto de-identification is the configuration method that uses AI to detect PII in a data source and de-identify it, without manual rule definition. It produces the same rule model as the other methods, a generator and mode bound to a field path but derives the rules from detection. It is one of three methods; the shared sequence and common reference tables are in De-identification Overview.
When to use it
- De-identifying a source quickly without enumerating fields.
- Detecting PII that may not be known in advance.
Key behaviors
AI detection drives the rules. Rules are produced by automatic PII detection rather than manual binding. The detected rules are then reviewable/ editable.
Same rule model. Detected rules resolve to the same generator, mode, and field-path bindings documented in De-identification Overview, so a field de-identified automatically is comparable to one configured manually.
Shares the common flow. After the auto step, filter, output, and summary are identical to the other methods.