| 2,056 | 5 | 331 |
| 下载次数 | 被引频次 | 阅读次数 |
人工智能仲裁裁决是仲裁法领域的最新发展。效率性、经济性和公正性的特性使得人工智能仲裁裁决具有可欲性,而人工智能算法由“规则驱动型”向“数据驱动型”的范式转变则提供了技术上的可行性。从合同法角度来看,人工智能仲裁裁决的执行实为仲裁协议下仲裁当事人的合同义务。在合同效力上,除了协议中指定的人工智能算法存在歧视这一违背程序性公序良俗事由外,人工智能仲裁协议应为有效;从仲裁法角度来看,人工智能仲裁裁决依然是《仲裁法》意义上的仲裁裁决。在裁决效力上,人工智能仲裁裁决仅可能因为算法歧视问题被认定为违背程序性社会公共利益而被撤销或不予执行。我国现有《仲裁法》以人类仲裁裁决为基本预设,应予更新以回应人工智能仲裁裁决的挑战。在基本思路上《,仲裁法》应当聚焦人工智能算法的特殊挑战,包括算法歧视、算法黑箱和算法裁决不合理等问题,同时平衡“公正与效率”“自治与干预”和“开放与谨慎”。在具体制度上《,仲裁法》应当构建人工智能仲裁裁决算法审查制度、赋予仲裁当事人算法解释权、扩展撤销和不予执行仲裁裁决法定事由。
Abstract:Arbitral awards made by artificial intelligence are the latest development in the field of arbitration law. The characteristics of efficiency, economy and fairness make arbitral awards made by artificial intelligence desirable, and the paradigm transformation of artificial intelligence from "rule-driven" to "data-driven" provides technical feasibility. From the perspective of contract law, the enforcement of arbitral awards made by artificial intelligence is actually the contractual obligations of the parties to the arbitration under the arbitration agreement. In terms of the validity of the contract, the artificial intelligence arbitration agreement shall be valid except that the artificial intelligence algorithm specified in the agreement is discriminatory, which is contrary to procedural public order and good morals. From the perspective of arbitration law, arbitral awards made by artificial intelligence are still arbitral awards in the sense of the Arbitration Law of the People's Republic of China. In terms of the validity of the awards, arbitral awards made by artificial intelligence can only be revoked or not implemented because the algorithm discrimination problem is found. China's existing arbitration law takes arbitral awards made by human as the basic presupposition and should be updated to respond to the challenge of arbitral awards made by artificial intelligence. As the basic guideline, the Arbitration Law of the People's Republic of China should focus on the special challenges of artificial intelligence algorithms, including algorithm discrimination, algorithm black box and unreasonable awards, while balancing "justice and efficiency", "autonomy and intervention" and "openness and prudence". As the specific norms,the Arbitration Law of the People's Republic of China should establish an algorithm review system, give the parties to the arbitration the right to interpret the algorithm, and expand the legal reasons for revocation and non-enforcement of the arbitral awards.
[1]朱体正.人工智能辅助刑事裁判的不确定性风险及其防范——美国威斯康星州诉卢米斯案的启示[J].浙江社会科学,2018(6):76-85+157.
[2]刘雁鹏.智慧司法中的忧虑:想象、剖析与展望[J].理论与改革,2020(3):168-181.
[3] Queen Mary University of London. 2021 International arbitration survey:adapting arbitration to a changing world[R/OL].(2021-05-06)[2021-12-08]. http://www. arbitration. qmul. ac. uk/media/arbitration/docs/2021-International-Arbitration-Survey-Adapting-arbitration-to-a-changing-world.pdf.
[4]深圳国际仲裁院.3i机器人再升级AI赋能仲裁服务[EB/OL].(2018-08-06)[2021-12-08].http://www.scia.com.cn/Home/Index/newsdetail/id/2670.html.
[5]张力行.创造与梦想[A]//深圳国际仲裁院.泉眼无声:国际仲裁的特区故事.北京:北京大学出版社,2020:152.
[6]江伟,李浩.论人民法院与仲裁机构的新型关系──为《仲裁法》的颁行而作[J].法学评论,1995(4):32-39.
[7]曼昆.经济学原理:微观经济学分册:第7版[M].梁小民,梁砾,译.北京:北京大学出版社,2015:291.
[8]冯姣,胡铭.智慧司法:实现司法公正的新路径及其局限[J].浙江社会科学,2018(6):67-75+85+157.
[9]王永庆.人工智能原理与方法[M].西安:西安交通大学出版社,1998:13.
[10] ALPAYDIN E. Introduction to machine learning[M].Cambridge:MIT Press,2014:332-333.
[11] SELA A. Can computers be fair:how automated and human-powered online dispute resolution affect procedural justice in mediation and arbitration[J].Ohio state journal on dispute resolution,2018,33(1):91-148.
[12]潘庸鲁.人工智能介入司法领域的价值与定位[J].探索与争鸣,2017(10):101-106.
[13]孙海波.反思智能化裁判的可能及限度[J].国家检察官学院学报,2020,28(5):80-99.
[14]罗素,诺维格.人工智能:一种现代的方法:第3版[M].殷建平,祝恩,刘越,译.北京:清华大学出版社,2013:606.
[15]陈海虹,黄彪,刘锋,等.机器学习原理及应用[M].成都:电子科技大学出版社,2017:2-14.
[16]俞勇.人工智能技术入门:让你也看懂的AI“内幕”[M].上海:上海科技教育出版社,2019:87.
[17]俞勇.从人脑到人工智能:带你探索AI的过去和未来[M].上海:上海科技教育出版社,2019:85.
[18] KATZ D M, BOMMARITO M JⅡ, BLACKMAN J. A general approach for predicting the behavior of the supreme court of the United States[J].Plos one,2017,12(4):1-18.
[19]高翔.人工智能民事司法应用的法律知识图谱构建——以要件事实型民事裁判论为基础[J].法制与社会发展,2018,24(6):66-80.
[20]瑟格·阿比特博,吉尔·多维克.算法小时代:从数学到生活的历变[M].任轶,译.北京:人民邮电出版社,2017:15-16.
[21]白建军.法律大数据时代裁判预测的可能与限度[J].探索与争鸣,2017(10):95-100.
[22]施米托夫.国际贸易法文选[M].赵秀文,选译.北京:中国大百科全书出版社,1993:598.
[23] RAYMOND A H, SHACKELFORD S J. Technology, ethics and access to justice should an algorithm be deciding your case[J].Michigan journal of international law,2014,35(3):485-524.
[24]韩健.现代国际商事仲裁法的理论与实践[M].北京:法律出版社,2000:36.
[25]乔欣.仲裁法学[M].北京:清华大学出版社,2015:72-74.
[26]宋朝武.仲裁法学[M].北京:北京大学出版社,2013.
[27]江伟,肖建国.仲裁法[M].北京:中国人民大学出版社,2016.
[28] SIM C. Will artificial intelligence take over arbitration[J].Asian international arbitration journal,2018,14(1):1-14.
[29]冉克平.论效力性强制规范与私法自治——兼析《民法总则》第153条第1款[J].山东大学学报(哲学社会科学版),2019(1):145-157.
[30]郑成良,牛安琪.强制性规定类型识别的辨伪与存真[J].社会科学战线,2021(5):215-225.
[31] LINDQUIST D H, DAUTAJ Y. AI in international arbitration:need for the human touch[J].Journal of dispute resolution,2021(1):39-64.
[32]李训虎.刑事司法人工智能的包容性规制[J].中国社会科学,2021(2):42-62+205.
[33]郑曦.人工智能技术在司法裁判中的运用及规制[J].中外法学,2020,32(3):674-696.
[34] SCHERER M. Artificial intelligence and legal decision-making:the wide open?[J].Journal of international arbitration,2019,36(5):539-574.
[35]何其生.国际商事仲裁司法审查中的公共政策[J].中国社会科学,2014(7):143-163+207-208.
[36]张兴伟.日本商事仲裁司法审查中的公共政策适用问题[J].国家行政学院学报,2016(5):131-135+145.
[37] EIDENMULLER H, VARESIS F. What is an arbitration? Artificial intelligence and the vanishing human arbitrator[J].New York university journal of law and business, 2020,17(1):49-94.
[38]伯恩·魏德士.法理学[M].丁晓春,吴越,译.北京:法律出版社,2013:309-310.
[39]雷震文.算法偏见对“智慧司法”的影响及其防范[N].法制日报,2017-12-27(11).
[40]魏斌.司法人工智能融入司法改革的难题与路径[J].现代法学,2021,43(3):3-23.
[41]卡尔·拉伦茨.法学方法论[M].黄家镇,译.北京:法律出版社,2020:508-518.
[42]约翰·罗尔斯.正义论[M].何怀宏,何包钢,廖申白,译.北京:中国社会科学出版社,1988:82.
[43]姜霞.仲裁司法审查程序本质论[J].河北法学,2007(6):154-157.
[44]哈贝马斯.在事实与规范之间:关于法律和民主法治国的商谈理论[M].童世骏,译.北京:生活·读书·新知三联书店,2011:280.
[45]罗伯特·阿列克西.法律证立理论——为法律证立理论的理性论辩理[M].舒国滢,译.北京:中国法制出版社,2002:366-373.
[46] KASAP G H. Can artificial intelligence("AI")replace human arbitrators? Technological concerns and legal implications[J].Journal of dispute resolution,2021(2):209-254.
[47]左卫民.从通用化走向专门化:反思中国司法人工智能的运用[J].法学论坛,2020,35(2):17-23.
[48]高鲁嘉.人工智能时代我国司法智慧化的机遇、挑战及发展路径[J].山东大学学报(哲学社会科学版),2019(3):115-123.
[49] MARROW P B, KAROL M, KUYAN S. Artificial intelligence and arbitration:the computer as an arbitrator-are we there yet?[J]. Dispute resolution journal, 2020,74(4):35-76.
[50]张凌寒.商业自动化决策的算法解释权研究[J].法律科学(西北政法大学学报),2018,36(3):65-74.
[51]张恩典.大数据时代的算法解释权:背景、逻辑与构造[J].法学论坛,2019,34(4):152-160.
[52]李飞.人工智能与司法的裁判及解释[J].法律科学(西北政法大学学报),2018,36(5):32-41.
(1)本文中引用的《仲裁法》皆系2017年修正版本。
(2)算法黑箱是指算法的运算逻辑处于未知状态,在数据输入和结果输出之间存在断层。
(3)算法歧视,也称算法偏见,是指那些可以造成不公平、不合理结果的系统性可重复出现的错误,其最常见的是对不同人有不同的结果,或者是给两个相同或相似条件的人不同结果。
(4)比如将《民法典》第577条转换成算法语句置入人工智能算法,遇到新的输入情形后,人工智能算法将首先判断输入的新情形是否满足“当事人一方不履行合同义务或者履行合同义务不符合约定”条件。具体而言,人工智能会对照设计者提前设定的违约的诸多具体情形一一进行检验,最后输出结论。
(1)同样以《民法典》第577条为例,在判断什么构成“当事人一方不履行合同义务或者履行合同义务不符合约定”情形时,“数据驱动型”人工智能并不寻求分析违约的精确定义也无须作出价值裁量(这正是“规则驱动型”人工智能采取的路径),而是通过对已有案例数据的分析,从文本发掘可能的影响认定违约的因素(即解释变量或自变量),并探索这些因素与违约认定结果(即被解释变量或因变量)之间的统计学相关性。
(2)合同内容主要是权利义务的规定,而权利和义务相对应,故只讨论义务一面即可。须指出,仲裁协议的内容还包括报酬事宜和仲裁当事人配合仲裁审理事宜等,但是此处只讨论最核心的内容。
(1)有例外即“该强制性规定不导致该民事法律行为无效的除外”。
(1)Interim Report on Public Policy as a Bar to Enforcement of International Arbitral Awards, International Law Association London Conference(2000).
(2)Final Report on Public Policy as a Bar to Enforcement of International Arbitral Awards, International Law Association New Delhi Conference(2002).
(1)当然除了违背程序正义外,基于仲裁的合意性,如果没有仲裁协议或超裁,司法也可干预仲裁裁决的效力。
基本信息:
DOI:10.15886/j.cnki.hnus.202109.2505
中图分类号:TP18;D925.7
引用信息:
[1]吴维锭.人工智能仲裁裁决:兴起、法律效力和法律回应[J].海南大学学报(人文社会科学版),2024,42(05):196-206.DOI:10.15886/j.cnki.hnus.202109.2505.
基金信息:
中国人民大学2021年度拔尖创新人才培育资助计划项目
2023-05-05
2023-05-05
2023-05-05