WHO AI for Health: Introduction & Six Topic Areas

Artificial intelligence (AI) — including machine learning and, increasingly, large language models — can strengthen health service delivery, public health surveillance, research, and clinical decision-making. Yet many AI tools are being deployed before their performance, limitations, and risks are fully understood, which can benefit or harm patients and the professionals who care for them.

The World Health Organization (WHO) and the International Telecommunication Union (ITU) established the Focus Group on AI for Health (FG-AI4H) and its Working Group on Regulatory Considerations (WG-RC) to increase international dialogue on emerging good practices. The 2023 publication Regulatory considerations on artificial intelligence for health summarises that work for developers, regulators, manufacturers, and health practitioners who deploy AI in care.

What this document is — and is not

WHO stresses that this overview is not formal regulatory guidance or policy. It is a resource of key considerations stakeholders may use when developing frameworks and best practices. Hong Kong healthcare professionals can use it to ask better questions of vendors, inform procurement, and align local practice with global expectations — alongside Hong Kong instruments such as the Personal Data (Privacy) Ordinance (PDPO) and the Medical Device Administrative Control System (MDACS) under the Department of Health.

Six topic areas

The WG-RC organised considerations into six areas (Table 1 in the source document):

1. Documentation and transparency
2. Risk management and AI systems development lifecycle approaches
3. Intended use and analytical and clinical validation
4. Data quality
5. Privacy and data protection
6. Engagement and collaboration

Across these areas, the working group highlights 18 recommendations (summarised in Table 5) covering pre-specification of development steps, total product lifecycle risk management, external validation, graded clinical evidence, data quality safeguards, privacy compliance programmes, and stakeholder engagement platforms.

Why this matters in Hong Kong

In public hospitals, private clinics, radiology, pathology, pharmacy, and allied health, AI is already embedded in workflows — from triage chatbots to imaging assistants and documentation tools. Responsible use requires understanding:

• whether a tool's intended use matches your patient population and setting;
• what documentation exists for training data and performance;
• how post-deployment monitoring will catch drift or failure; and
• how privacy and cybersecurity risks are managed under local law.

The following articles in this learning path unpack each WHO topic area in plain language, with short quizzes to reinforce AI literacy for daily practice.

Source: WHO — Regulatory considerations on artificial intelligence for health (2023)