Meta: course=ai-legal-governance · module=1 · lesson=1.1 · ~85 min · keywords: WHO six principles, human autonomy, well-being and safety, transparency, accountability, inclusiveness equity, sustainability, African Union AI strategy, national AI strategies, regulatory sandbox Objectives:
- State the WHO six principles for AI ethics and governance and explain each in a health context.
- Describe the consolidating African policy layer — AU continental strategy, national strategies, SaMD pathways.
- Explain how regulatory sandboxes let evidence and rules co-develop.
WHO's guidance on AI ethics and governance provides the consensus frame the whole field aligns to: six principles (Ch7 §7.6.2). They are: (1) protecting human autonomy; (2) promoting human well-being and safety; (3) transparency and explainability; (4) responsibility and accountability; (5) inclusiveness and equity; and (6) responsive and sustainable AI. Each maps onto operational practice. Human autonomy underwrites the augmentation frame and informed consent — the patient and clinician retain agency; the tool advises. Well-being and safety demands local validation and post-deployment monitoring. Transparency and explainability is the deployment checklist's explainability item. Responsibility and accountability is the "responsible institution" requirement — an institution answers, never the algorithm. Inclusiveness and equity is the bias and benefit-sharing agenda. Responsiveness and sustainability asks whether a deployment is maintainable, affordable, and environmentally and institutionally durable — not a pilot that collapses when the grant ends. These are not abstractions; they are the principles the legal and institutional instruments of this course operationalize.
The African policy layer is now consolidating around these principles, and an SME must be able to locate a deployment within it. The book reports that the African Union's continental AI strategy (2024) sets the harmonization direction, national AI strategies are multiplying, and medicines and device regulators are extending software-as-a-medical-device (SaMD) pathways to algorithmic tools (Ch7 §7.6.2). The direction of travel is clear: AI in health is moving from an ungoverned space into a regulated one, harmonized continentally and implemented nationally. This matters commercially and operationally — interoperability and compliance are "market access" (Ch4), and an AI tool that ignores the emerging policy layer is building toward a wall.
The book singles out one instrument as the practical venue where the policy layer actually develops: the regulatory sandbox. Sandboxes "are the practical venue where evidence and rules co-develop" (Ch7 §7.6.2), allowing innovations — AI triage, drone corridors, novel insurance — "to operate under supervision, with defined scope and monitoring, while evidence and rules co-develop" (Ch4 §4.5.1). The sandbox solves a real chicken-and-egg problem: a regulator cannot write good rules for a category it has never observed, and an innovator cannot generate evidence for a category that is prohibited by ambiguity. The sandbox converts the regulator "from gatekeeper to learning partner without abandoning safety" — and converts the innovation from prohibited-by-ambiguity into supervised evidence generation. For AI specifically, the sandbox is where TB-CAD-adjacent triage tools, conversational assistants, and predictive models can be observed under real conditions before general rules harden around them.
The strategic point for African ownership recurs here. The book frames the choice sharply: Africa "as data source and deployment market for intelligence owned elsewhere — or Africa as builder, validator, and governor of intelligence" (Ch7 §7.7). A policy layer that merely imports foreign approvals makes Africa a deployment market; a policy layer that validates and governs — local validation requirements, African datasets under African governance, AI-literate regulators in functioning sandboxes — builds African ownership. The six principles are universal; the institutions that enact them must be African, and building them is, the book insists, the highest-return investment in the entire stack (Ch4 §4.5.2).
Key terms:
- WHO six principles — protecting human autonomy; promoting human well-being and safety; transparency and explainability; responsibility and accountability; inclusiveness and equity; and responsive and sustainable AI.
- Software as a Medical Device (SaMD) — software intended for a medical purpose that functions without being part of a hardware device; the regulatory category many health AI tools fall under.
- Regulatory sandbox — a supervised environment where innovations operate under defined scope and monitoring while evidence and regulation co-develop, turning the regulator from gatekeeper to learning partner.
- AU Continental AI Strategy (2024) — the African Union's framework setting the harmonization direction for national AI strategies and health-AI regulation.
Knowledge check: Q: Name the WHO six principles for AI ethics and governance. A: Protecting human autonomy; promoting human well-being and safety; transparency and explainability; responsibility and accountability; inclusiveness and equity; and responsive and sustainable AI.
Q: What chicken-and-egg problem does a regulatory sandbox solve? A: Regulators cannot write good rules for categories they have never observed, and innovators cannot generate evidence for categories prohibited by ambiguity; the sandbox lets evidence and rules co-develop under supervision.
Q: How does the policy layer determine whether Africa is a "deployment market" or an "owner"? A: Merely importing foreign approvals makes Africa a deployment market; requiring local validation, building African datasets under African governance, and resourcing AI-literate regulators in functioning sandboxes builds ownership.
Summary: WHO's six principles are the consensus frame, each mapping to operational practice. Africa's policy layer is consolidating — AU continental strategy, national strategies, SaMD pathways — with the regulatory sandbox as the venue where evidence and rules co-develop. Whether the layer imports approvals or validates and governs decides whether Africa is a deployment market or an owner of its health AI.