The European Commission’s final report, Study on the deployment of AI in healthcare, released in July–August 2025, offers a rigorous mixed-methods evaluation of the challenges and accelerators affecting the integration of Artificial Intelligence (AI) into clinical practice across EU Member States and peer nations. Drawing upon literature reviews and stakeholder consultations conducted between January 2024 and January 2025, the study identifies systemic obstacles—technological, regulatory, organisational, and sociocultural, while illustrating successful strategies and proposing a monitoring and indicators framework to guide sustainable, ethical scale-up of AI in healthcare systems.
Healthcare systems in the European Union are grappling with escalating challenges: an aging population, growing prevalence of chronic diseases, and workforce shortages. Between 2000 and 2023, the share of the population aged 65 and above rose from 16% to over 21%, projected to reach nearly 30% by 2050. Concurrently, 40% of these older adults live with at least two chronic conditions. AI presents a potential path forward—enhancing efficiency, diagnostic precision, and care delivery—but practical deployment in clinical settings remains limited.
The study aimed to delimit the most pressing challenges and enabling conditions for AI deployment in clinical practice. It examines experiences across EU Member States and selected third countries—specifically, the USA, Japan, and Israel—with advanced AI integration in healthcare. The research employed a mixed-methods approach, coupling literature reviews with stakeholder consultations. It identifies major obstacles to deployment of the AI in healthcare, but also promising practices, and offers recommendations.
We bring you the key findings:
- Barriers to AI Deployment
1. Technological & Data Limitations
o Interoperability and data standardisation remain insufficient, impeding seamless AI integration and scalability.
o Variable quality and fragmented datasets hinder model training and generalisability.
2. Legal & Regulatory Complexity
o Navigating a complex landscape—such as GDPR and evolving medical device directives—presents compliance burdens.
o The AI-specific regulations in development further compound these challenges.
3. Organisational & Business Issues
o Healthcare institutions often lack the infrastructure and resource capacity to pilot and scale AI systems.
o Financial incentives for adoption and staff training are insufficient.
4. Social & Cultural Resistance
o Concerns over AI’s reliability and transparency fuel mistrust among clinicians and patients.
o Integration into existing clinical workflows remains a hurdle.
- Accelerators and Promising Practices
1. International Exemplars: Cases from the USA, Japan, and Israel demonstrate feasible strategies for overcoming AI adoption barriers.
2. Innovative Models: Hospitals' use of AI for administrative optimisation, diagnostics, or patient monitoring offers practical blueprints.
3. Stakeholder Engagement: Active co-design with clinicians and patients fosters greater trust and usability.
- Recommnedations for Effective AI Deployment in Healthcare:
1. Patient-Centric Design
AI tools must prioritize patient safety, privacy, and equity. Design algorithms with explainability and transparency in mind.
2. Data Governance & Quality
Standardize and ensure high-quality, representative datasets. Establish strong data-sharing frameworks with privacy safeguards.
3. Regulatory Alignment
Collaborate with health authorities to develop dynamic, clear AI regulatory pathways. Encourage post-deployment monitoring and audits.
4. Workforce Integration & Training
Train clinicians and staff on AI capabilities and limitations. Position AI as an assistive tool, not a replacement.
5. Interoperability & Infrastructure
Invest in digital infrastructure that supports seamless integration with existing health IT systems. Promote open standards and cross-platform compatibility.
6. Continuous Evaluation
Implement real-world testing environments and feedback loops.
A novel framework is proposed to track deployment progress, enabling policymakers and healthcare administrators to evaluate AI uptake, effectiveness, and equity metrics across time and regions. Strategic or normative guidance is particularly valuable in this transformative period. The report aligns with broader EU initiatives like the AI Act entering into force in August 2024 and expected full applicability by August 2026.
