GDPR Article 9 — Special Category (Health) Data — Regulatory Reference

GDPR Article 9(1) prohibits processing of special category data — including health records — by default. For AI agents, this means every retrieval of health data must be tied to a lawful basis under Article 9(2): explicit consent, vital interests, healthcare provision, or public health. An agent that queries patient data without a verified lawful basis is in violation, regardless of whether a human reviewed the output. AutoPIL enforces this at the retrieval layer: consent state is propagated into access policy, and any agent request without an active, matching consent record is denied before the data enters the context window.

GDPR Article 9 applies whenever an AI agent processes data that reveals health status, diagnoses, treatment history, genetic data, or biometric data used for identification — and the data subject is in the EU or EEA, or the controller is established in the EU. This includes clinical AI tools, diagnostic assistants, care coordination agents, and population health analytics platforms. It applies regardless of where the AI infrastructure is hosted. Organizations outside the EU that process EU patient data as part of a service offering are also in scope under GDPR's extraterritorial reach in Article 3.

GDPR Article 83(5) sets the maximum fine for Article 9 violations at €20 million or 4% of global annual turnover, whichever is higher. Supervisory authorities have issued significant fines for health data misuse: the Swedish DPA fined a health region for unauthorized access by staff; the Spanish AEPD has repeatedly fined hospitals for inadequate access controls. For AI systems, regulators increasingly expect technical enforcement — logging that an agent ‘should have’ respected consent is not sufficient. Demonstrable pre-retrieval enforcement and an audit trail of every access decision are becoming baseline expectations in supervisory guidance.

Article 17 gives data subjects the right to request deletion of their health records, including any derived data. For AI systems, a key compliance gap is proving that agents no longer have access to erased records — standard deletion from source systems does not prevent retrieval from cached context or vector stores. AutoPIL’s tamper-evident audit chain enumerates every AI access to a data subject’s records by agent ID, timestamp, and data source. When a right-to-erasure request is received, this audit history confirms what was accessed and when, supporting deletion verification across all data pathways. Policy HC-GDPR-A17-001 is the reference configuration for access history scoping.

GDPR Chapter V (Articles 44–49) restricts transfers of health data to third countries without adequate protection or a valid transfer mechanism such as Standard Contractual Clauses or a Binding Corporate Agreement. For AI agents deployed across cloud regions, this creates a real enforcement gap: an agent querying a patient record may inadvertently route the retrieval through infrastructure in a non-adequate country. AutoPIL policy HC-GDPR-CHV-001 enforces cross-border transfer rules at the data access layer — the source registry records data residency, and agents are denied access to sources outside their permitted transfer scope before any data is retrieved.