Data Minimisation

Data minimisation is the principle that an AI agent should access and process only the minimum data necessary to complete its task. It is a core requirement under GDPR and a key defence against both accidental data leakage and deliberate misuse. In agent governance, minimisation is enforced through scoped credentials, fine-grained access policies, and output filtering.