生成式 AI醫學應用之倫理考量

本文分析生成式AI(GAI)醫學應用之倫理考量，首先介紹生成式AI與傳統式AI之差異，分析GAI可能帶來的衝擊，於醫學應用之趨勢，進而探討GAI於醫療應用時之倫理考量，包括「資料倫理」層面應關注：隱私、資料保護、資料所有權、二次使用、可解釋性、真實性、偏見等議題，於「臨床倫理」層面則由：醫療適應症、病人偏好、生活品質、情境特徵4主題方法進一步探討。作者指出GAI在醫學領域應用，必須兼顧「資料倫理」與「臨床倫理」兩個層面方能確保其可課責性與可信賴性。

This paper examines the ethical considerations surrounding the medical applications of Generative AI (GAI). It first delineates the distinctions between GAI and traditional AI, analyses the potential impacts of GAI, and explores emerging trends in its medical applications. The discussion then addresses key ethical concerns related to GAI in healthcare, structured across two dimensions: Data Ethics–encompassing issues of privacy, data protection, ownership, secondary use, interpretability, authenticity, and bias. Clinical Ethics–explored through the four-topic methods: medical indications, patient preferences, quality of life, and contextual features. The authors argue that the responsible and trustworthy integration of GAI in medicine necessitates a dual focus on both data ethics and clinical ethics, ensuring accountability, transparency, and ethical integrity in its application.