Inference & Profiling · Restriction or exception
AI Products: no automated decisions in domains that affect material or individual rights or well-being
Making automated decisions in domains that affect material or individual rights or well-being (for example finance, law, healthcare, housing, insurance and social welfare) is unacceptable use of AI models and AI-powered Products. This item sits alongside the Section 3.2(b) restriction on automated eligibility determinations for credit, employment, education or public assistance.
Supporting quotations
(h) Automated decision-making in sensitive domains. This means making automated decisions in domains that affect material or individual rights or well-being (e.g. finance, law, healthcare, housing, insurance, and social welfare).
Qualifications & exceptions
- Applies to AI models and Products powered by AI models (Section 3.3).
- The domains are introduced with "e.g.", so the list is not stated to be complete.
Not stated here
- The document does not say how this item and the narrower Section 3.2 restrictions are meant to interact beyond listing both.