Abstract Background/Introduction Breathlessness evaluation remains challenging for clinicians, with differential diagnosis between cardiac and non-cardiac causes often requiring extensive testing. Current diagnostic approaches achieve variable accuracy, with emergency department heart failure diagnosis showing sensitivity of 65-66% and specificity of 90-92% using traditional methods 1. Existing artificial intelligence applications in dyspnoea evaluation lack systematic ethical frameworks, leading to technically proficient but ethically deficient deployments. Isaac Asimov's Three Laws of Robotics provide foundational principles for safe AI implementation, emphasising patient safety and system integrity preservation. Purpose To develop the CURACO-Asimov framework for ethical AI implementation in breathlessness evaluation, incorporating Asimov's safety-first principles with systematic integration of Clinical safety, Understanding, Research-informed care, Authentic patient-centred approaches, Conscientious ethics, and Optimised technology. We aimed to demonstrate framework applicability for AI-enhanced diagnostic tools differentiating cardiac versus non-cardiac breathlessness causes. Methods We developed the CURACO-Asimov framework by integrating Asimov's Three Laws with six core principles: safety-first diagnostic protocols, patient understanding enhancement, evidence-based implementation, authentic patient-centred approaches, conscientious ethics integration, and human-AI collaboration optimisation. The framework was applied to analyse current AI applications in dyspnoea evaluation, including diagnostic algorithms, biomarker interpretation systems, and imaging analysis tools. Results The CURACO-Asimov framework identified critical ethical implementation gaps in current AI systems for breathlessness evaluation. Framework application revealed significant diagnostic accuracy variations: AI models for dyspnoea achieved 96.5% overall diagnostic accuracy 1, while traditional BNP testing shows sensitivity of 94.1% and specificity of 74.5% for cardiac causes. Bedside ultrasonography demonstrated superior performance with area under ROC curve of 86.4% compared to BNP at 66.3% for cardiogenic dyspnoea detection 2. Despite these clinical capabilities, implementation barriers persist including limited algorithmic transparency and inadequate patient empowerment components. Conclusion The CURACO-Asimov framework provides a novel systematic approach for ethical AI implementation in breathlessness evaluation. By integrating Asimov's safety-first principles with comprehensive ethical considerations, the framework transforms technical AI capabilities into patient-centred, ethically robust diagnostic tools. While this initial framework development demonstrates broad applicability, empirical validation studies with quantitative metrics are needed.
Sooltan et al. (Thu,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: