WHO warns Europe on healthcare AI: deployment is outpacing safety rules

The World Health Organization’s 2024-2025 survey says Europe is rolling out healthcare AI faster than it’s putting patient safeguards in place.

Just 8% of the 53 countries in the WHO European Region have a health-specific AI strategy, the survey found, even though nearly two-thirds already use AI in diagnostics and about half have patient-facing chatbots.

That same 2024-2025 WHO survey found that 85% of EU states have a broad AI strategy, but when it comes to healthcare, the planning is much thinner.

WHO says fewer than half of those countries have checked whether their current health laws are actually fit for purpose, and nearly 40% still don’t have ethical guidance. Hans Kluge says that leaves room for biased diagnoses, clinicians using tools they can’t explain, and a mess over who’s accountable when systems fail.

Training is thin, too: WHO says about 20% of countries teach AI during initial training, and about 25% offer in-service training, which means staff may be working with software they don’t fully understand.

The upside is real as well. WHO says 98% of states report better patient care, and in Coimbra, WHO points to image analysis that’s helping speed up thoracic disease and fracture detection while cutting wait times.

If you work in healthcare or policy, this should get your attention.

You can read the full WHO findings for the rest of the picture, including the EU AI Act timeline, the July 2026 Lisbon meeting with 37 countries, and a 2028 Lusophone roadmap.