Technical audit of healthcare AI: what the data controller must verify - Agenda Digitale

▤ Summary

This article explains what a data controller must verify when a healthcare software product uses AI. It goes beyond contract checks and calls for scrutiny of the model’s decision logic, training data provenance, robustness, monitoring, and human oversight. It also highlights GDPR risks tied to special-category health data, including bias, traceability, and possible leakage from training data. The piece concludes that a proper DPIA should rest on technical evidence, not commercial documentation alone.

▦ Linked Companies