AI 101

Anonymization

Also known as: anonymise · de-identification

Security

Anonymization removes or transforms identifiers so data can’t be re-linked to a person.

  • Replace names, emails, IDs with tokens.
  • Generalize specifics (city → region).
  • A step before feeding data to models or sharing it.

The caveat: true anonymization is hard — quasi-identifiers can re-identify. For strong guarantees, pair it with pseudonymization and access controls.