with the introduction of the AI, such as bias and social inequality, and has institutionalised protective measures in the form of patient rights. In constract patients in Japan tend to take a more passive role, and institutional responses to these concerns appear weaker. This comparison suggests that while the UK adopts a patient-centered ethical approach, Japan’s approach remains more focused on the perspectives of healthcare providers and institutions.academic and policy discussions. This concept emphasises that the successful integration of AI in healthcare requires more than technological performance. It must also align with the cultural, institutional, and social context in which it is deployed. For instance, even if an AI system provides accurate recommendations, it may not be effectively utilized unless it fits seamlessly into physicians’daily workflows. Additionally, cultural considerations and ergonomic knowledge are also required for decisions such as the format of the proposal and who will explain it to the patient. This perspective captures the introduction of AI not merely as a technical challenge, but as a sociotechnical one requiring the design of human-AI interactions that respect professional roles, patient relationships, and societal values. This approach to implementation design comprehensively addresses the alignment ofAI with healthcare systems, ethical principles, and the perspectives of both healthcare professionals and patients. As an example, the Institute for AI and Medical Relationships in Japan conducts research on the ethical and institutional challenges surrounding AI adoption, while also developing models for social implementation. Its approach to integrating medical AI is multifaceted, encompassing technology, policy, and culture. In summary, both countries agree that medical AI’s performance itself is not enough to implement into real-world settings, but they take different approaches. The UK focuses on a field-based approach, whereas Japan has a comprehensive and socially oriented perspective. Furthermore, the major disadvantages of implementing AI such as false negatives (missed diagnosis), false positives (misdiagnosis), and inappropriate interventions are highlighted. In addition to conserns about accuracy, the potential for the AI to cause incorrect decisions in clinical settings is explicity addressed as well. It is also started that the AI-driven technologies could disproportionately disadvantage socially vulnerable groups, such as those in poverty or ethnic minorities, with the risk of increasing social disparities. Therefore, the UK openly recognises the risks associated The second difference is that there are institutional and cultural differences in how concerns are addressed. While both countries agree that technical performance alone is insufficient, they diverge in how AI implementation is evaluated. The UK prioritizes empirical clinical utility, whereas Japan places greater empgasis on cultural and institutional alignment. These differences reflect distinct interpretations of what constitutes‘successful implementation’in each context. Firstly, the BMA highlights that the AI is not about how accurately it performs in laboratory settings, but how useful in real-world clinical settings, arguing that most research projects over-focus on the accuracy of the AI in lab environments, where conditions are ideal. For example, a Google-developed AI system for detecting diseases performed well during laboratory testing, however, it did not function well in actual settings. This is because this AI is trained by high-quality images with a camera, the quality of images in real-world clinical situations is so low that the AI did not respond to images that did not meet the standards. Although the primary aim was to alleviate staff shortages, the AI system ended up with increasing staff workload instead. This example highlights the need for AI to be designed with a clear understanding of its use in real-world settings. While the BMA acknowledges that very few AI models perform effectively both in the laboratory consitions and in practice, efforts are underway to address this challenge, including the introduction of empirical evaluation systems.In contrast, Japan has yet to develop an institutional framework that explicitly addresses the practical implementation of AI in clinical settings, as seen in the UK. Instead, the concept of “cultural implementation” proposed by Fujita of the Tokyo Foundation, has gained traction in 73. Implementation Evaluation
元のページ ../index.html#8