Trust, Risk and Governance in AI-Driven Healthcare Systems: A Behavioral and Policy Perspective

The responsible integration of Artificial Intelligence (AI) in healthcare depends on understanding the behavioural, organisational, and regulatory factors shaping clinicians’ and organisations’ trust in AI-enabled Clinical Decision Support Systems (AI-CDSSs). This chapter conducts a multi-level empirical investigation combining three levels of analysis. First, a behavioural survey of 215 clinicians from Italy and the United Kingdom was analysed using Structural Equation Modelling (SEM) grounded in Prospect Theory and the UTAUT model. Second, a patentometric network analysis examined over 8,000 patent families (2014-2024) from international patent databases. Third, a computational and qualitative discourse analysis assessed 95 European Union (EU) policy documents (2018-2025). Findings show that Trust mediates the relationship between Perceived Risks and Behavioural Adoption Intention (β = −0.46, p < 0.001), with loss aversion the primary barrier to adoption. AI-healthcare innovation is structurally concentrated among a small network of influential developers, with China and the United States accounting for over 80% of patent production. The policy analysis reveals a measurable shift in EU discourse from ethics-driven governance (2018-2021) to compliance-based regulation (2022-2025). Together, these results suggest that trust cannot be treated as a mere compliance mechanism within the established legal framework, particularly in data science and machine learning. Instead, trust must be cultivated through transparency, algorithmic explainability, and adaptable regulatory processes. The chapter proposes a multi-level governance framework for the responsible adoption of AI in healthcare, with practical implications for policy design, clinical procurement, and institutional leadership.