AIRiskAware
AI Governance Glossary
Governance Concept

What Is Anonymisation?

Anonymisation is processing personal data so that individuals can no longer be identified from it, irreversibly and in a way that cannot reasonably be undone.

Definition

Anonymisationprocessing personal data so that individuals can no longer be identified from it, irreversibly and in a way that cannot reasonably be undone.

Truly anonymised data falls outside data-protection law, because there is no longer an identifiable person to protect. The bar is high: if data can be re-identified with reasonable effort — by combining it with other available information — it is not anonymous but merely pseudonymised, and the obligations still apply. Getting this distinction wrong is a common compliance trap in AI projects.

Source: GDPR, Recital 26; ICO guidance

Plain-language explanation

Truly anonymised data falls outside data-protection law, because there is no longer an identifiable person to protect. The bar is high: if data can be re-identified with reasonable effort — by combining it with other available information — it is not anonymous but merely pseudonymised, and the obligations still apply. Getting this distinction wrong is a common compliance trap in AI projects.

Primary source: GDPR, Recital 26; ICO guidance

Related terms

Pseudonymisation Re-identification Data Minimisation Differential Privacy

See where you stand on AI governance

Take the free 7-question maturity assessment and get a personalised action plan.

Free assessment — 3 minutes →