GDPR Document Anonymization Template Prompt
Before a document can be shared, analysed or used to train anything, the personal data inside it has to be dealt with. This prompt identifies what qualifies as personal data under the GDPR, distinguishes direct identifiers from indirect ones, and proposes how to remove or pseudonymize each without making the document useless.
How to Use the Prompt:
- The document to anonymize: paste the text. Contracts, reports, transcripts and correspondence all carry more personal data than they appear to.
- The purpose: anonymization for publication, for internal sharing and for statistical analysis require different levels of aggressiveness.
- What must remain readable: state which information the document has to preserve to still serve its purpose.
Benefits of Using this Prompt:
- Indirect identifiers are caught: a role, a date and a location together can identify a person as precisely as a name.
- Consistent pseudonyms: the same person keeps the same label across the document, so the text remains coherent.
- Usability is preserved: blanking everything is easy and makes the document worthless; the point is removing what identifies, not what informs.
- A reasoned position: the output explains why each element was treated the way it was, which is what an audit asks for.
- A repeatable procedure: the same criteria apply to the next document, and the one after.
There is a limit that no prompt gets around: to be anonymized by an external service, the document has to leave your organisation first. For personal data, that is often precisely what you are not allowed to do.