-
摘要: 医疗人工智能中的算法黑箱在机器的实践中造成了突出的伦理问题与风险,算法黑箱这一概念的本质是人类对人机关系失控的形象化描述,医疗人工智能算法黑箱的核心伦理问题在于人机关系的失控。关于问题的解决,本文在反思以透明化为代表的技术叠加方案的基础上,提出解决问题的核心策略必须立足于机器所面向的医疗实践,超越技术视域的局限性,寻求有意义的控制,使机器的实践回归医疗人工智能技术发展的根本目标,更好地满足现实医疗需求和促进人类健康福祉。Abstract: The algorithm black box in medical artificial intelligence(AI) creates prominent ethical problems and risks in machine practices. The essence of the concept of the algorithm black box is a figurative description of human beings' loss of control in human-machine relationship, and the core ethical issue of the algorithm black box in medical AI lies in the loss of control. Regarding the solution to the problem, based on reflecting on the technology overlay solution represented by transparency, this paper proposes that the core strategy for solving the problem must be based on the medical practice that the machine is oriented towards, transcending the limitations of the technical perspective, seeking meaningful control, so that the practice of the machine can return to the fundamental goal of the development of medical AI technology to meet the reality of medical needs and promote the well-being of human health.
-
-
[1] 周文康, 费艳颖. 医疗人工智能前沿研究: 特征、趋势以及规制[J]. 医学与哲学, 2021, 42(19): 38-40.
[2] 浮 婷. 算法“黑箱”与算法责任机制研究[D]. 北京: 中国社会科学院研究生院, 2020: 9.
[3] 陶迎春. 技术中的知识问题: 技术黑箱[J]. 科协论坛: 下半月, 2008(7): 54-55.
[4] SMITH H. Clinical AI: Opacity, accountability, responsibility and liability[J]. AI Soc, 2021, 36(2): 535-545. doi: 10.1007/s00146-020-01019-6
[5] 斯坦纳. 算法帝国[M]. 李筱莹, 译. 北京: 人民邮电出版社, 2014: 42-43.
[6] 徐 恪, 李 沁. 算法统治世界: 智能经济的隐形秩序[M]. 北京: 清华大学出版社, 2017: 11.
[7] BURRELL J. How the machine 'thinks': Understanding opacity in machine learning algorithms[J]. Big Data Soc, 2016, 3(1): 205395171562251.
[8] 李春生. 技术治理中的算法“黑箱”及其应对策略[N]. 中国社会科学报, 2021-11-10(008).
[9] 科技部. 关于加快场景创新以人工智能高水平应用促进经济高质量发展的指导意见[EB/OL]. (2022-07-29)[2024-03-19]. https://www.gov.cn/zhengce/zhengceku/2022-08/12/content_5705154.htm.
[10] 李 明, 李昱熙, 戴 廉, 等. 医疗人工智能伦理若干问题探讨[J]. 医学与哲学, 2019, 40(21): 1-4. doi: 10.12014/j.issn.1002-0772.2019.21.01
[11] 彭运朋, 徐毅华. 医疗人工智能对知情同意的挑战与应对[J]. 医学与哲学, 2023, 44(10): 25-29. doi: 10.12014/j.issn.1002-0772.2023.10.06
[12] 徐着雨, 岳远雷. 医疗人工智能算法风险防范的法治化思考[J]. 医学与哲学, 2023, 44(11): 67-71. doi: 10.12014/j.issn.1002-0772.2023.11.15
[13] QUINN T P, JACOBS S, SENADEERA M, et al. The three ghosts of medical AI: Can the black-box present deliver?[J]. Artif Intell Med, 2022, 124: 102158. doi: 10.1016/j.artmed.2021.102158
[14] 李润生. 论医疗人工智能“黑箱” 难题的应对策略和规制进路[J]. 东南大学学报: 哲学社会科学版, 2021, 23(6): 83-92.
[15] 张姝艳, 皮婷婷. 医疗领域中人工智能应用的可解释性困境与治理[J]. 医学与哲学, 2023, 44(3): 25-29. doi: 10.12014/j.issn.1002-0772.2023.03.06
[16] HATHERLEY J J. Limits of trust in medical AI[J]. J Med Ethics, 2020, 46(7): 478-481. doi: 10.1136/medethics-2019-105935
[17] 段伟文. 控制的危机与人工智能的未来情境[J]. 探索与争鸣, 2017(10): 7-10. doi: 10.3969/j.issn.1004-2229.2017.10.002
[18] The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems. Ethically Aligned Design: A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems-Version II[EB/OL]. (2017-12-12)[2024-03-19]. https://standards.ieee.org/wp-content/uploads/import/documents/other/ead_v2.pdf.
The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems. Ethically Aligned Design: A Vision for Prioritizing Human Well-being with Autonomous and Intelligent Systems-Version II[EB/OL]. (2017-12-12)[2024-03-19]. https://standards.ieee.org/wp-content/uploads/import/documents/other/ead_v2.pdf.
[19] World Health Organization. Ethics and governance of artificial intelligence for health[EB/OL]. (2021-09-28)[2024-03-19]. https://www.who.int/publications/i/item/9789240029200.
[20] 国家新一代人工智能治理专业委员会. 新一代人工智能伦理规范[EB/OL]. (2021-09-26)[2024-03-19]. https://www.safea.gov.cn/kjbgz/202109/t20210926_177063.htm.
[21] 张 荣, 徐 飞. 人工智能医学伦理问题及对策研究[J]. 医学与哲学, 2020, 41(13): 14-15.
[22] 鲍 曼. 现代性与矛盾性[M]. 邵迎生, 译. 北京: 商务印书馆, 2003: 288.
[23] 亚里士多德. 尼各马可伦理学[M]. 廖申白, 译注. 北京: 商务印书馆, 2003: 3.
[24] 温纳著. 自主性技术作为政治思想主题的失控技术[M]. 杨海燕, 译. 北京: 北京大学出版社, 2014: 21.
[25] 刘永谋, 王春丽. 智能时代的人机关系: 走向技术控制的选择论[J]. 全球传媒学刊, 2023, 10(3): 5-21.
[26] 王 硕, 文侃骁, 刘天语, 等. 医疗机器人的伦理风险与治理探析: 基于“医−患−机” 分析框架[J]. 医学与哲学, 2023, 44(23): 16-21.
[27] 国家人工智能标准化总体组 全国信标委人工智能分委. 人工智能伦理治理标准化指南: 2023[EB/OL]. (2023-04-12)[2024-03-19]. https://nits.org.cn/index/article/4133.
[28] 纪守领, 李进锋, 杜天宇, 等. 机器学习模型可解释性方法、应用与安全研究综述[J]. 计算机研究与发展, 2019, 56(10): 2071-2096.
[29] 谭九生, 范晓韵. 算法“黑箱” 的成因、风险及其治理[J]. 湖南科技大学学报: 社会科学版, 2020, 23(6): 92-99.
[30] 沈伟伟. 算法透明原则的迷思: 算法规制理论的批判[J]. 环球法律评论, 2019, 41(6): 20-39.
[31] 吕乃基. 论科技黑箱[J]. 自然辩证法研究, 2001, 17(12): 23-26. doi: 10.3969/j.issn.1000-8934.2001.12.006
[32] DURÁN J M, JONGSMA K R. Who is afraid of black box algorithms?: On the epistemological and ethical basis of trust in medical AI[J]. Journal of Medical Ethics, 2021, 47: 334.
计量
- 文章访问数: 2118
- PDF下载数: 1259
- 施引文献: 0