Early Online.

Abstract

Artificial intelligence (AI) is rapidly reshaping clinical practice, and has prompted the Health Professions Council of South Africa to publish Booklet 20 - its first ethical framework for AI use in healthcare. This review critically evaluates Booklet 20 through the lens of the South African Medical Association AI Task Team, examining its alignment with national legislation, emerging regulatory mechanisms, broader policy commitments and the ethico-social context. Drawing on the Protection of Personal Information Act, the South African Health Products Regulatory Authority's 2025 guidance for AI‑enabled medical devices, the National AI Policy Framework and the 2024 National Health Research Ethics Council ethics guidelines, the analysis identifies key operational gaps relating to human oversight, disclosure, data sovereignty, equity, accountability and risk categorisation. The article argues that while Booklet 20 establishes an important foundation, its principles require concrete implementation tools, including risk‑tiered safeguards, structured consent templates, meaningful governance co-ordination and context‑appropriate standards for transparency, explainability and bias mitigation. Grounding these enhancements in South Africa's communitarian ethic of ubuntu highlights the need for relational accountability, fairness and community participation to ensure safe and equitable AI integration. The article concludes with a set of practical recommendations aimed at strengthening ethical governance and supporting patient trust and professional integrity as AI becomes embedded in clinical workflows.