Learn
Articles and quizzes on healthcare AI, regulation, and responsible innovation.
Follow the curated paths below — designed for Hong Kong clinicians, pharmacists, and healthcare administrators building responsible AI literacy step by step.
Curated learning paths
Three guided tracks based on EMA guidance, from everyday LLM use to medicines lifecycle regulation.
Start here: Responsible AI in daily practice
Build foundational AI literacy with EMA's network strategy and practical guidance for using large language models safely in clinical, pharmacy, and administrative work.
- Step 1EMA's AI Strategy for Medicines RegulationHow the European medicines regulatory network plans to harness AI while managing risks — and what Hong Kong healthcare professionals can learn from its 2025–2028 workplan.Read article →
- Step 2Using Large Language Models Responsibly in HealthcareEMA's four guiding principles for safe LLM use — adapted for Hong Kong clinicians, pharmacists, and administrators who draft, summarise, or research with AI tools.Read article →
International standards & regulatory milestones
Understand joint EMA–FDA principles for AI in medicines development and how regulators are validating AI-assisted evidence.
- Step 1EMA & FDA: Ten Principles for Good AI PracticeJoint international principles for using AI across the medicines lifecycle — from early research and clinical trials to manufacturing and safety monitoring.Read article →
- Step 2AI Regulatory Milestones: Qualification & ObservatoryFrom EMA's first AI methodology qualification (AIM-NASH) to the 2024 AI Observatory — what these milestones mean for evidence standards and AI literacy.Read article →
Deep dive: AI across the medicines lifecycle
Work through EMA's reflection paper in order — from general principles and lifecycle stages to regulatory interactions, technical expectations, and governance.
- Step 1EMA AI Reflection Paper: Introduction & General PrinciplesHow the European Medicines Agency frames AI and machine learning in medicines regulation, and the risk-based principles that apply across the lifecycle.Read article →
- Step 2AI Across the Medicinal Product LifecycleEMA guidance on using AI from drug discovery and clinical trials through manufacturing, product information, and post-authorisation pharmacovigilance.Read article →
- Step 3Regulatory Interactions for AI in MedicinesWhen and how to engage EMA through the Innovation Task Force, scientific advice, and qualification of novel methodologies for AI/ML.Read article →
- Step 4Technical Aspects of AI/ML for MedicinesEMA expectations for data acquisition, model development, performance assessment, explainability, and deployment of AI in regulated settings.Read article →
- Step 5Governance, Data Protection & Trustworthy AIEMA guidance on GxP governance, GDPR compliance, data integrity, and EU trustworthy AI principles for medicinal products.Read article →
WHO guidance on large multi-modal models in health
Build AI literacy for generative AI in clinical care — from LMM basics and clinical risks to WHO's six ethical principles, governance, and responsible deployment.
- Step 1WHO Guidance on Large Multi-Modal Models: IntroductionWhat LMMs are, why they differ from earlier clinical AI, and how WHO's 2024 guidance builds on six consensus ethical principles for Hong Kong healthcare professionals.Read article →
- Step 2LMMs in Diagnosis, Patient Care & AdministrationPotential benefits and major risks when LMMs support clinical decision-making, patient-facing chatbots, and clerical tasks — with practical checks for Hong Kong practice.Read article →
- Step 3LMMs in Education, Research & Health System RisksFrom medical student chatbots to scientific writing — plus systemic risks of overestimating LMM benefits, digital divides, and health workforce impacts.Read article →
- Step 4Regulatory Compliance & Societal Risks of LMMsData protection, medical device boundaries, big-tech dominance, environmental footprints, and threats to human epistemic authority.Read article →
- Step 5WHO's Six Ethical Principles for LMMs in PracticeApply autonomy, well-being, transparency, accountability, equity, and sustainability when evaluating LMM tools in Hong Kong healthcare.Read article →
- Step 6Governance, Deployment & Responsible Use of LMMsNavigate the AI value chain — development, provision, and deployment — with deployer duties, workforce training, and liability considerations for Hong Kong settings.Read article →
IMDRF Good Machine Learning Practice for medical devices
Work through IMDRF's ten guiding principles for AI-enabled medical devices — from intended use and data validation to human–AI teamwork, transparency, and real-world monitoring.
- Step 1IMDRF Good Machine Learning Practice: IntroductionWhy international regulators published ten guiding principles for AI-enabled medical devices, and what Hong Kong healthcare professionals should know before using them in practice.Read article →
- Step 2GMLP Principles 1–2: Intended Use & Engineering FoundationsDefine clinically meaningful intended use with multidisciplinary teams, and implement robust software engineering, security, and quality practices across the device lifecycle.Read article →
- Step 3GMLP Principles 3–5: Data, Validation & Reference StandardsEnsure training and test data represent your patient population, maintain independence between datasets, and use fit-for-purpose reference standards.Read article →
- Step 4GMLP Principles 6–7: Model Design & Human–AI TeamsMatch model design to clinical purpose and evaluate performance of the human–AI team in real workflow — not the algorithm alone.Read article →
- Step 5GMLP Principles 8–10: Testing, Transparency & MonitoringValidate AI devices under clinically relevant conditions, provide clear information to users, and monitor deployed models for drift and retraining risks.Read article →
MHRA: Software and AI as a medical device (UK)
Learn how the UK's MHRA regulates software and AI medical devices — from the change programme and post-market vigilance to international PCCP and transparency principles.
- Step 1MHRA Software & AI as a Medical Device: IntroductionHow the UK's MHRA regulates software as a medical device (SaMD) and AI as a medical device (AIaMD), and why Hong Kong healthcare professionals should understand this framework when evaluating UK-origin or internationally harmonised products.Read article →
- Step 2MHRA Change Programme: SaMD Lifecycle & AI Work PackagesThe eleven work packages across two streams — SaMD lifecycle reforms and AI-specific challenges — including qualification, classification, post-market vigilance, and Yellow Card reporting.Read article →
- Step 3MHRA PCCP Guiding Principles for ML-Enabled DevicesThe five joint MHRA/FDA/Health Canada principles for predetermined change control plans — focused, risk-based, evidence-based, transparent, and total product lifecycle — with links to IMDRF GMLP.Read article →
- Step 4MHRA Transparency Principles for ML-Enabled DevicesThe who, why, what, where, when, and how framework for transparency in machine learning medical devices, human-centred design, and links to GMLP principles 7 and 9.Read article →
WHO: Regulatory considerations on AI for health
Build AI literacy with WHO's six topic areas — documentation, lifecycle risk, validation, data quality, privacy, and stakeholder engagement — for responsible use in Hong Kong practice.
- Step 1WHO AI for Health: Introduction & Six Topic AreasWhy WHO and ITU established global dialogue on AI in health, what rapid deployment means for Hong Kong clinicians, and how six regulatory topic areas support responsible AI literacy.Read article →
- Step 2Documentation & Transparency Across the AI LifecyclePre-specify development plans, trace decisions and dataset choices, and use proportional documentation to build trust — lessons from WHO for Hong Kong AI adopters.Read article →
- Step 3Risk Management & the AI Systems LifecycleApply a total product lifecycle approach — pre-market, post-market surveillance, and change management — with holistic risk controls for cybersecurity, bias, and model updates.Read article →
- Step 4Intended Use, Analytical & Clinical ValidationDefine use cases clearly, validate externally on representative populations, and match clinical evidence to risk — from retrospective metrics to trials and post-deployment monitoring.Read article →
- Step 5Data Quality for Safe & Effective Health AIAssess whether data are fit for purpose, run rigorous pre-release checks, and mitigate bias, labelling errors, and representativeness gaps before AI reaches patients.Read article →
- Step 6Privacy & Data Protection in Health AIUnderstand jurisdictional privacy laws, design privacy from the start, and separate cybersecurity incidents from broader privacy harms — with PDPO and cross-border transfer awareness for Hong Kong.Read article →
- Step 7Engagement, Collaboration & the Way ForwardBuild stakeholder platforms, position regulators as partners, and sustain dialogue so AI innovations reach patients safely — lessons for Hong Kong institutions and professional bodies.Read article →
More articles
Additional learning materials outside the curated paths.
FDA AI Device TPLC: Validation, Monitoring, Transparency & Cybersecurity
Performance validation, postmarket monitoring, transparency and bias strategies, and AI-specific cybersecurity risks from FDA's draft TPLC guidance.
Read article →FDA AI Device TPLC: Data, Models & Validation Terminology
FDA's draft expectations for data management, model description and development, and the critical terminology differences between FDA 'validation' and AI community usage.
Read article →FDA AI Device TPLC: Submissions — Description, UI & Risk
What FDA expects in marketing submissions for device description, user interface and labeling, and risk assessment — and what Hong Kong clinicians should look for in vendor documentation.
Read article →FDA AI Device TPLC Guidance: Introduction
How FDA's draft total product lifecycle guidance applies to AI-enabled device software functions, and what Hong Kong clinicians should know about scope, GMLP, and predetermined change control plans.
Read article →Good AI Practice Principles 8–10: Assurance
Conduct risk-based performance assessment including human–AI interactions, manage AI through the full life cycle, and communicate clear essential information to users and patients.
Read article →Good AI Practice Principles 5–7: Development
Integrate multidisciplinary expertise, maintain rigorous data governance and documentation, and follow sound model design practices for reliable AI in drug development.
Read article →Good AI Practice Principles 1–4: Foundation
Human-centric design, risk-based validation, adherence to GxP and other standards, and a clear context of use — the foundational principles for AI in drug development.
Read article →Good AI Practice in Drug Development: Introduction
Why FDA and EMA published ten guiding principles for AI in drug development, what scope they cover, and what Hong Kong clinicians, pharmacists, and researchers should know.
Read article →Guiding Principles of Good AI Practice in Drug Development
U.S. FDA guidance on responsible AI use across the drug product life cycle (January 2026).
Read article →