What changed in October 2026?
On 29 October, the 53 countries in WHO’s European Region adopted a shared roadmap for AI in health for 2026–2033. It focuses on governance, safeguards, reliable digital foundations and practical implementation. The region includes countries in Europe and Central Asia.
This is a plan for responsible health-system development, not an approval of every AI app. It does not establish that a particular chatbot can diagnose illness or that a new tool is available in your local clinic.
Sources: WHO/Europe — New roadmap for AI in health, 29 October 2026
A medical AI tool and a chatbot are different
Some medical devices use AI for a defined task, such as helping interpret medical images. Their intended use, evidence and regulatory status matter. In the United States, the FDA evaluates medical devices through relevant regulatory pathways; an authorization concerns the particular device and use, not AI technology as a whole.
A general chat assistant generates language. It may help you prepare questions or understand a concept, but a fluent answer is not a physical examination, a verified test interpretation or a clinical diagnosis.
Sources: FDA — Artificial intelligence-enabled medical devices; WHO — Benefits and risks of generative AI in health
Where it may help, and where errors matter
WHO describes possible uses including administrative summaries, health education, research and support for clinical care. Whether a tool improves outcomes depends on how well it is tested and used. A result in one setting does not automatically carry over to another.
Generative AI can invent facts or references, miss important context and produce biased or incomplete information. Training data may poorly represent some ages, languages or communities. A source link should be opened and checked: its presence alone does not prove that the claim is supported.
Sources: WHO — Benefits and risks of generative AI in health; WHO — Safety, validation and regulation of health AI
A practical way to use AI for health education
- Ask for a plain-language explanation or questions to take to an appointment.
- Check important claims against an established health agency or your healthcare team.
- Keep names, identification numbers, medical documents and other identifying details out of general chat services.
- Read how a service handles information before sharing anything sensitive.
- Ask who checks an AI tool’s output and what happens when it is uncertain.
For example, “What questions could I ask about a cholesterol test?” is a useful starting point. “Tell me which medicine to stop” asks for a personal treatment decision that belongs with a qualified prescriber.
Sources: WHO — Safety, validation and regulation of health AI; WHO — Benefits and risks of generative AI in health
Keep a route back to human care
Ask a clinician using an AI tool what it is designed to do, whether it has been evaluated for people like you, and how its result affects the next step. You can also ask what alternatives are available if you prefer to discuss a concern directly.
Do not change prescribed treatment, dismiss worsening symptoms or delay urgent care because of a chat response. For an emergency, contact your local emergency service rather than waiting for AI. HealthBook AI is an optional educational assistant, not a medical device or a diagnostic service; the same limits apply to it.
Sources: WHO — Safety, validation and regulation of health AI; FDA — Artificial intelligence-enabled medical devices
Common questions
Does a detailed AI answer mean it is medically correct?
No. Detail and confidence are not evidence of accuracy. Confirm important health information with reliable sources and a qualified professional.
Does the new WHO roadmap approve HealthBook AI?
No. It is a regional policy roadmap, not an endorsement of HealthBook or an approval of individual chat services.
Sources & further reading
Public sources consulted for this update. Guidance can differ by country and personal circumstances.
HealthBook provides general health information for educational purposes and is not a substitute for professional medical advice, diagnosis, or treatment.