Loading ...

Although artificial intelligence (AI) tools are increasingly available, their integration into clinical practice remains slow, according to a new European Commission report. Barriers include data and technology gaps, legal and regulatory complexity, organisational and financial constraints, and social and cultural resistance.

The final report of the Study on the Deployment of AI in Healthcare analyses the development and use of AI throughout the sector, identifying both the main challenges and the factors that can accelerate implementation.

The report also compiles international best practices, highlighting strategies hospitals have used to address these barriers, and introduces a monitoring framework with indicators to track progress toward sustainable integration of AI in healthcare.

Data Trust

Data fragmentation prevents the effective integration of AI into clinical workflows, and outdated infrastructure cannot support the use of modern applications.

The report also notes the absence of standardised evaluation protocols and the lack of sustained postmarket monitoring, which undermines confidence among clinicians.

This challenge is compounded by the “black box” phenomenon, in which the inner workings of an algorithm, particularly those on the basis of deep learning or complex neural networks, are difficult to understand.

Another driver of mistrust among healthcare professionals and patients is limited digital health literacy and technological competence, which hampers the understanding of how AI tools function, their potential and limitations, and their role as support tools in healthcare delivery.

Legal Barriers

The study reveals that the EU has a robust and complex legal and regulatory landscape. Interactions between multiple regulations, combined with privacy, security, and liability concerns, contribute to cautious implementation.

Financial barriers are another challenge, with limited reimbursement or funding mechanisms for AI systems. Insufficient involvement of end users during development often leads to tools that are poorly aligned with clinical needs, and the lack of standardised models to evaluate local added value hinders rigorous assessment.

Global Lessons

The report draws on the experiences of countries that have advanced AI adoption in healthcare, including the US, Israel, and Japan. Their health systems have applied a range of best practices to address these challenges, which the study concludes could be further explored in the European context to support wider implementation.

Furthermore, the EU is uniquely positioned to support the safe, effective, ethical, and equitable scaling of AI in healthcare by balancing innovation and protecting patients’ fundamental rights. By doing so, the EU can establish itself as a global leader in AI-driven healthcare and deliver accessible, high-quality, and sustainable care.

Future Actions

The report outlines several priority steps to overcome current barriers and enable sustainable AI adoption in European healthcare.

• Data standards: Common frameworks for data governance and interoperability across health systems should be established to enable seamless AI integration, secure cross-border exchanges, and regulatory compliance.
• Centres of excellence: Create hubs to address skill gaps, provide advanced training for health professionals, promote digital literacy, and foster collaboration in AI innovation.
• Funding mechanisms: Develop consolidated financing models to support AI projects and ensure equitable access to tools across health systems.
• Evaluation frameworks: Value-added assessments, local performance studies through assurance laboratories, and sustained postimplementation monitoring should be introduced to guarantee effectiveness, safety, and compliance.
• AI catalogue: Build a central EU repository of available AI solutions to guide informed decisions and drive innovation across the EU.

This story was translated from El Medico Interactivo.