Hans Kluge, WHO Regional Director for Europe, told delegates in Lisbon on 15 July that while artificial intelligence is already being rolled out across health systems in the region, governance and workforce preparation have not kept pace. He opened his remarks with one striking figure: “8 percent” — that is the share of countries in the WHO European Region that have a dedicated national strategy for health AI.
Findings of the WHO survey
The WHO surveyed its 53 member states in the European Region and described the exercise as its most comprehensive assessment to date of AI preparedness in health. The survey found that nearly two‑thirds of countries already use AI for diagnostic purposes, and half of member states have introduced AI‑based patient communication chatbots. Despite broad adoption, only one in twelve countries has developed a comprehensive governance strategy for health AI.
The detailed results point to wider gaps in training and legal preparedness:
- Only one‑fifth of countries provide AI‑related education to health professionals before they qualify.
- For existing health workers, just one in four countries offers AI‑related continuing professional development.
- Fewer than half of member states have reviewed whether their legal frameworks are suitable for managing AI deployment.
- Nearly 40 percent have no ethical guidance at all for the use of AI in health.
Kluge illustrated the risks with concrete consequences: a biased algorithm can produce a wrong diagnosis for a real patient, and a clinician trained to trust an AI system whose functioning they cannot question or verify is not empowered to make appropriate decisions — leading to errors beyond their control.
The issue is governance, not the technology itself
Kluge emphasised that the problem is not the AI per se. Ninety‑eight percent of WHO member states view improving patient care as the primary aim of health AI. He cited Coimbra, Portugal, where AI‑assisted image analysis helps clinicians detect chest disease and bone fractures more quickly, reducing waiting times in primary and emergency care. Real patients are already receiving better care thanks to AI, he said. The concern is that the technology is being used in settings where responsibility and accountability for failures have not been clearly established.
Recommendations and initiatives
Kluge set out three main tasks:
- Regulatory frameworks must keep pace with deployment: every country using health AI needs a strategy, clear rules on responsibility and adequate workforce training.
- International cooperation must be strengthened: to this end the WHO convened representatives from 37 countries in Lisbon from all six WHO regions.
- Portuguese‑speaking countries should have a specific collaborative role: Portugal, Angola, Brazil, Mozambique and partners are preparing a Portuguese‑language cooperation roadmap on AI and health, intended to be presented at the WHO’s 2028 Regional Health Summit in Brazil.
The Lisbon conference and priority topics
The conference, organised with the Portuguese government and opened by Health Minister Ana Paula Martins, ran through 16 July. It focused on three strands: AI regulation and accountability, tools needed for safe implementation, and preparing the workforce who will use the technology in everyday practice.
Kluge noted that Europe already has institutions and infrastructure for this work: the European Union has established testing centres that support areas such as medical imaging and robotic rehabilitation. The central problem, he argued, is a gap between these developments and the 53 European health ministries, because most countries have yet to develop the basic regulatory frameworks.
Concluding point
Kluge warned that the future of health AI will be shaped not by algorithms themselves but by the regulatory frameworks created now, the partnerships established, and the political will to ensure this technology serves everyone rather than only those countries and communities rich enough to define its path.
