AI readiness in European healthcare: WHO report 2025

Chatbots and virtual assistants

Across half of the WHO European Region, conversational AI tools now play an active role in addressing medication queries, scheduling appointments, providing symptom guidance and supporting service navigation.

By offering immediate, accessible support, these virtual assistants ease pressure on frontline teams, streamline routine interactions and improve the overall patient experience through faster responses and clearer guidance.

Automating administrative and clerical tasks

High administrative burden is a shared challenge across European health systems. AI is now supporting:

• Scheduling
• Coding
• Documentation
• Logistics
• Workflow optimisation.

These tools help free clinicians to spend more time with patients.

AI-assisted symptom checkers and decision support

An increasing number of countries are deploying AI-powered symptom checkers and clinical decision-support tools to assist with patient triage, guide diagnostic reasoning and inform treatment planning.

These systems analyse patient-reported information and clinical data to provide timely insights, helping clinicians prioritise cases, reduce uncertainty and enhance the safety and consistency of decision-making across busy care environments.

Prognostic tools and risk stratification

AI-powered prognostic models are increasingly being used to identify patients at risk of deterioration far earlier than traditional methods allow.

By analysing patterns across clinical records, imaging and laboratory data, these tools support more proactive interventions, help clinicians prioritise high-risk cases and improve the timeliness and safety of care delivery.

Robotics and AI-assisted surgery

AI-enabled robotic systems are becoming increasingly established in surgical practice, particularly in countries such as France and Spain. These technologies enhance precision, support complex procedures and improve consistency in operating theatres.

By assisting surgeons with real-time guidance, motion stabilisation and automated workflow support, AI-driven robotics are helping to standardise outcomes and reduce variation in surgical care.

Remote patient monitoring

AI-enabled monitoring is being used for chronic disease management, early warning systems and post-operative follow-up.

These use cases align with systemwide pressures across Europe, where workforce shortages, rising demand and increasing complexity require new forms of digital augmentation.

Northern Europe: The most advanced overall

Countries including Finland, Sweden, Norway, Denmark and Estonia show:

• Strong national AI or digital health strategies
• Robust data governance frameworks
• Widespread health data hubs
• Strong cross-border data exchange mechanisms
• Early adoption of AI diagnostics, triage tools and administrative automation
• The strongest workforce development programmes.

Northern Europe also leads in both pre-service and in-service AI training (WHO report p. 18).

Western Europe: Strong in governance, private-sector partnerships and high-end use cases

Countries such as France, Belgium, the Netherlands and Switzerland excel in:

• Diagnostic AI
• Robotic surgery
• Chatbots
• AI-enabled clinical decision tools
• Public–private partnerships
• Regulatory oversight.

Central and Eastern Europe: High innovation in specific use cases

Countries such as Hungary and Slovakia are emerging as early innovators, particularly in diagnostic imaging and radiotherapy, where AI offers clear clinical gains and addresses workforce shortages.

Despite having less mature national strategies or digital infrastructure, these systems demonstrate agility by rapidly piloting and scaling targeted AI solutions that respond to immediate clinical and operational pressures.

Southern Europe: Rapid adoption but governance gaps

Countries like Spain and Portugal show strong ambition, especially in:

• Robotic surgery
• Chatbots
• Administrative automation.

But they lag behind Northern and Western Europe in workforce training and governance maturity.

Legal uncertainty: The number one barrier

Member States identified legal ambiguity as the greatest obstacle:

• Unclear liability
• Uncertainty around AI classification
• Lack of clear standards
• Minimal post-market monitoring.

This is shown clearly in the WHO’s barrier chart (p. 47).

Financial constraints

Financial barriers continue to slow progress, with many health systems struggling to fund the infrastructure, licensing and subscription models needed to deploy AI at scale.

Ongoing costs, such as system maintenance, upgrades and workforce training, further stretch limited budgets. For smaller or resource-constrained countries, these financial pressures can delay adoption or restrict AI use to isolated pilots rather than systemwide implementation.

Weak data governance and interoperability

Even though:

• 66% have national health data strategies
• 76% have some governance framework
• 66% have national/regional health data hubs.

Only:

• 30% have cross-border data sharing rules
• 40% have frameworks for public–private data sharing (WHO report p. 34–38).

This limits the accuracy, fairness and safety of AI models.

Workforce gaps

Only 24% of countries provide in-service AI training, and just 20% offer pre-service education, leaving large sections of the health workforce underprepared for AI-enabled care.

This creates significant risks in safety-critical environments, where clinicians must understand system limitations, interpret outputs appropriately and maintain clinical judgement. Without targeted training, AI adoption may increase misuse, automation bias and variability in care quality.

Trust and ethical concerns

Trust in AI remains fragile, particularly when clinicians, patients and the wider public feel excluded from decision-making about new technologies. Limited transparency, unclear accountability and concerns about data use can heighten scepticism and resistance.

Without meaningful engagement, ethical reassurance and clear communication, confidence in AI systems weakens, ultimately slowing adoption and undermining their potential to enhance safe, person-centred care.