Trust Is the Bottleneck

Artificial intelligence (AI) is transforming health care across diagnostic imaging, clinical decision support, and personalized medicine. However, despite rapid technical advances, trust remains the principal bottleneck limiting clinical adoption. This report synthesizes insights from the second IEEE EMBS Workshop on AI and Healthcare, which centered on the multidimensional nature of trust in medical AI systems. Through keynote demonstrations of AI-enabled ultrasound and an interdisciplinary panel discussion spanning medicine, computational oncology, defense research, and scientific publishing, the workshop examined trust as a construct grounded not only in accuracy, but also in transparency, accountability, and interpretability. Audience polling and case-based discussions revealed that public and clinical confidence in AI is conditional and context-dependent, particularly in high-stakes decision environments. The workshop highlighted key tensions between performance and explainability, emphasized the need for explicit accountability frameworks, and underscored the importance of lifecycle oversight from data provenance to post-deployment validation. As AI becomes increasingly embedded in clinical workflows, its impact will depend less on algorithmic capability alone and more on its ability to earn and sustain trust. The discussions reinforce a critical conclusion: trustworthy healthcare AI requires validation, human-centered design, interdisciplinary collaboration, and sustained governance to evolve from promising technology to a reliable clinical partner.