Fu-Chang Tsai

Daniel Fu-Chang Tsai is a family physician and bioethicist. He graduated from the National Taiwan University College of Medicine (NTUCM) in 1989 and earned his PhD in bioethics from the University of Manchester, U.K. in 1999. He is the founding professor of the Graduate Institute of Medical Education & Bioethics, a joint professor in the Department of Family Medicine and the Graduate Institute of Clinical Medicine, and the past Director of the Department of Social Medicine at the NTUCM. He is an attending physician in the Department of Medical Research, the founding Director of the Ethics Center, the chairman of the Research Ethics Committee, and the executive secretary of the Clinical Ethics Committee at National Taiwan University Hospital. He is also the Director of the Center for Biomedical Ethics at National Taiwan University, the past Vice President of the International Association of Bioethics, and the past President of the Taiwan Association of Institutional Review Boards. He is a member of the Merk Ethics Advisory Panel. He has led many government-commissioned projects by the Ministry of Health & Welfare, the Ministry of Education, the Ministry of Science & Technology, and the National Health Research Institute. He has more than 230 papers published in Chinese & English journals, and 40 book chapters. He is on the editorial board of the Journal of Medical Ethics, Kennedy Institute of Ethics Journal, Asian Bioethics Review, and Formosa Journal of Medicine. He was awarded Honorary Membership by the UNESCO Chair of Bioethics in 2015 and elected Vice President of the International Association of Bioethics in 2016, the Goldman-Berland Lectureship in Palliative Medicine in 2019, elected as Hastings Center Fellow in 2021, and Fellow of the Chinese Association for General Education (2024). His special research interests include cross-cultural bioethics, genetic ethics, transplantation ethics, clinical ethics and ethics consultation, research ethics, and medical ethics education.

17th August 2025

Time	Session

10:30 12:00	Ethics/Law (Mandarin) 倫理與法律: 人工智慧衝擊醫療衍生的倫理與糾紛如何應對 (中文) 人工智慧衝擊醫療衍生的倫理與糾紛如何應對 Chung-You TsaiTaiwan Moderator Bridging AI Frontiers and Urology: How Multimodal and Agentic AI Will Shape 20251. Evolution of AI: From LLM to Agentic AI AI has progressed rapidly from basic language models (LLMs) to multimodal and agentic systems capable of autonomous decision-making and task execution. 2. General vs. Domain-Specific LLMs General-purpose LLMs offer versatility, while domain-specific LLMs (e.g., medical models) provide higher accuracy in specialized fields like urology. 3. AI Applications in Medical Practice LLMs and AI agents assist in research, academic writing, and clinical decision-making—transforming how urologists access and apply medical knowledge. 4. Agentic AI & Multi-Agent Systems AI agents can orchestrate tools, reason through complex problems, and automate workflows without human input—enhancing productivity in healthcare. 5. Benchmarking AI vs. Human Experts In prostate cancer risk assessment, top-tier LLMs demonstrated competitive or superior performance compared to human experts, indicating clinical potential. How to Make AI as the Most Powerful Assistance for the Treatment of GU Cancer? Hsiu-Nan TsaiTaiwan Moderator AI 醫療應用的倫理問題與挑戰 Fu-Chang TsaiTaiwan Speaker AI 醫療應用的倫理問題與挑戰人工智慧（AI）是當今科技發展的主流與大趨勢，其範疇幾乎無所不在，所將帶給人類的影響亦將是全面、本質性且不可逆轉。本演講將探討AI於醫院、醫療照護、醫學研究等應用發展現況，並從個人資訊的隱私保護與知情同意、資料管理與使用、建立社會信任三方面來分析其所涉倫理議題，並將進一步探討生成式AI於醫學研究與應用所衍生倫理法律問題。智慧醫療衍生的醫療糾紛—人工智慧誤診怎麼辦？該如何看待與應對？ Kai-Hsin ChangTaiwan Speaker 智慧醫療衍生的醫療糾紛—人工智慧誤診怎麼辦？該如何看待與應對？本研究報告深入分析了人工智慧（AI）在醫療診斷中潛在的誤診與誤判問題，並從技術成因、臨床影響、法律責任歸屬及監管應對等多面向進行了探討。本報告發現，AI誤診的根源在於其「黑箱」特性、訓練數據的偏差與不足，以及模型可能隨時間發生的性能退化。在臨床實踐中，儘管AI被定位為輔助工具，但其介入對醫師的判斷力帶來了新的挑戰，並持續重塑醫療照護的標準。現行侵權法原則（如醫療過失、轉承責任、產品責任）在AI情境下適用困難，尤其在因果關係證明以及AI軟體「產品」與「服務」的法律界定上存在模糊性。為應對這些挑戰，本文擬提出下列建議： 1. 增強技術穩健性與數據品質：強調使用多樣化、高品質的「黃金標準」數據集，並實施持續監測與反饋循環，同時優先發展可解釋AI（XAI）技術，以提升模型透明度和可理解性。 2. 強化人類監督與培訓：明確AI應作為輔助工具，並強制對醫療專業人員進行全面的AI應用培訓，建立清晰的內部治理政策與跨學科AI委員會，並實施健全的文檔記錄實踐。 3. 發展健全的法律與監管框架：比較美國FDA與歐盟《AI法案》等不同監管模式，指出歐盟採取更為全面的立法途徑，將醫療AI歸類為「高風險」系統，並透過產品責任指令（PLD）與AI責任指令（AILD）減輕受害者舉證責任。報告建議立法應明確各方責任，並參考國際經驗平衡創新與安全。 4. 推動保險解決方案的演進：分析現有醫療專業責任保險（MPLI）在AI時代面臨的覆蓋範圍空白與除外條款問題，建議保險業應開發AI特定保險產品，並調整承保趨勢以適應AI整合帶來的風險變化。本研究強調，醫療AI的負責任部署需透過技術、培訓、法規和保險等多層次、協同一致的策略，方能平衡創新潛力與患者安全，確保AI真正為人類健康福祉服務。討論 TICC - 2F 201AF

Time

Session

10:30

12:00

Ethics/Law (Mandarin)

倫理與法律: 人工智慧衝擊醫療衍生的倫理與糾紛如何應對 (中文)

人工智慧衝擊醫療衍生的倫理與糾紛如何應對

Chung-You TsaiTaiwan Moderator Bridging AI Frontiers and Urology: How Multimodal and Agentic AI Will Shape 20251. **Evolution of AI: From LLM to Agentic AI** AI has progressed rapidly from basic language models (LLMs) to multimodal and agentic systems capable of autonomous decision-making and task execution. 2. **General vs. Domain-Specific LLMs** General-purpose LLMs offer versatility, while domain-specific LLMs (e.g., medical models) provide higher accuracy in specialized fields like urology. 3. **AI Applications in Medical Practice** LLMs and AI agents assist in research, academic writing, and clinical decision-making—transforming how urologists access and apply medical knowledge. 4. **Agentic AI & Multi-Agent Systems** AI agents can orchestrate tools, reason through complex problems, and automate workflows without human input—enhancing productivity in healthcare. 5. **Benchmarking AI vs. Human Experts** In prostate cancer risk assessment, top-tier LLMs demonstrated competitive or superior performance compared to human experts, indicating clinical potential. How to Make AI as the Most Powerful Assistance for the Treatment of GU Cancer?

Hsiu-Nan TsaiTaiwan Moderator
AI 醫療應用的倫理問題與挑戰

Fu-Chang TsaiTaiwan Speaker AI 醫療應用的倫理問題與挑戰人工智慧（AI）是當今科技發展的主流與大趨勢，其範疇幾乎無所不在，所將帶給人類的影響亦將是全面、本質性且不可逆轉。本演講將探討AI於醫院、醫療照護、醫學研究等應用發展現況，並從個人資訊的隱私保護與知情同意、資料管理與使用、建立社會信任三方面來分析其所涉倫理議題，並將進一步探討生成式AI於醫學研究與應用所衍生倫理法律問題。
智慧醫療衍生的醫療糾紛—人工智慧誤診怎麼辦？該如何看待與應對？

Kai-Hsin ChangTaiwan Speaker 智慧醫療衍生的醫療糾紛—人工智慧誤診怎麼辦？該如何看待與應對？本研究報告深入分析了人工智慧（AI）在醫療診斷中潛在的誤診與誤判問題，並從技術成因、臨床影響、法律責任歸屬及監管應對等多面向進行了探討。本報告發現，AI誤診的根源在於其「黑箱」特性、訓練數據的偏差與不足，以及模型可能隨時間發生的性能退化。在臨床實踐中，儘管AI被定位為輔助工具，但其介入對醫師的判斷力帶來了新的挑戰，並持續重塑醫療照護的標準。現行侵權法原則（如醫療過失、轉承責任、產品責任）在AI情境下適用困難，尤其在因果關係證明以及AI軟體「產品」與「服務」的法律界定上存在模糊性。為應對這些挑戰，本文擬提出下列建議： 1. 增強技術穩健性與數據品質：強調使用多樣化、高品質的「黃金標準」數據集，並實施持續監測與反饋循環，同時優先發展可解釋AI（XAI）技術，以提升模型透明度和可理解性。 2. 強化人類監督與培訓：明確AI應作為輔助工具，並強制對醫療專業人員進行全面的AI應用培訓，建立清晰的內部治理政策與跨學科AI委員會，並實施健全的文檔記錄實踐。 3. 發展健全的法律與監管框架：比較美國FDA與歐盟《AI法案》等不同監管模式，指出歐盟採取更為全面的立法途徑，將醫療AI歸類為「高風險」系統，並透過產品責任指令（PLD）與AI責任指令（AILD）減輕受害者舉證責任。報告建議立法應明確各方責任，並參考國際經驗平衡創新與安全。 4. 推動保險解決方案的演進：分析現有醫療專業責任保險（MPLI）在AI時代面臨的覆蓋範圍空白與除外條款問題，建議保險業應開發AI特定保險產品，並調整承保趨勢以適應AI整合帶來的風險變化。本研究強調，醫療AI的負責任部署需透過技術、培訓、法規和保險等多層次、協同一致的策略，方能平衡創新潛力與患者安全，確保AI真正為人類健康福祉服務。
討論

TICC - 2F 201AF