生成式人工智能创作物的侵权责任界定探析——以ChatGPT为例
An Analysis on the Definition of Tort Liability for Generative Artificial Intelligence Creations—A Case Study of ChatGPT
摘要: 生成式人工智能这个火热的话题以ChatGPT的问世推向高潮,随即所面临新颖的侵权方式,是对现行法律体系的挑战,传统侵权法是难以对其侵权责任的承担做出应对。对此本文将通过分析生成式人工智能自身特点以及侵权行为特点即生成式人工智能创作内容的自主性、可信度和侵权行为隐蔽性的角度出发,围绕其特点对不同生成式人工智能主体理论学说做出价值判断即支持折中说,认为应赋予生成式人工智能有限的法律人格;再者根据其特点探讨出明确的判断标准即“接触 + 实质性相似”以解决生成式人工智能创作物侵权认定问题;最后以生成式人工智能创作物类型的划分、致人损害和虚构信息导致侵权三种侵权方式,确定生成式人工智能所划分的两种创作物类型即高技术性、专业性内容和思想观念表达内容分别适用过严格责任原则和无过错原则;而致人损害导致侵权适用严格责任原则更为合理以及虚构信息导致侵权可适用一般侵权责任原则加以解决。以此为解决侵权责任的界定提供参考价值。
Abstract: The hot topic of generative artificial intelligence reached its climax with the advent of ChatGPT, and then faced with novel infringement methods, which is a challenge to the current legal system, and the traditional tort law is difficult to cope with its infringement liability. In this regard, by analyzing the characteristics of generative AI and the characteristics of tort, namely the autonomy, credibility and concealability of generative AI’s creative content, this paper will make a value judgment on dif-ferent generative AI’s subjective theories, that is, support the compromise theory, and hold that generative AI should be endowed with limited legal personality. Furthermore, according to its characteristics, a clear judgment standard, namely “contact + substantial similarity”, is discussed to solve the problem of infringement identification of generative artificial intelligence creations. Finally, based on the classification of generative artificial intelligence creation types, causing damage to people and causing infringement by fictional information, it is determined that the two types of generative artificial intelligence creation types, namely high technical content, professional content and ideological expression content, have respectively applied the principle of strict liability and the principle of no-fault. It is more reasonable to apply the principle of strict liability for infringement caused by damage to human beings, and the general principle of tort liability can be applied to solve the problem caused by fictitious information. It provides reference value for solving the definition of tort liability.
文章引用:秦俊龙. 生成式人工智能创作物的侵权责任界定探析——以ChatGPT为例[J]. 法学, 2023, 11(6): 5046-5051. https://doi.org/10.12677/OJLS.2023.116722

参考文献

[1] 徐伟. 论生成式人工智能服务提供者的法律地位及其责任——以ChatGPT为例[J]. 法律科学(西北政法大学学报), 2023, 41(4): 69-80. [Google Scholar] [CrossRef
[2] 邓腾. 人工智能产品的侵权责任界定问题[J]. 中阿科技论坛(中英文), 2023(4): 141-145.
[3] 周书环. 聊天机器人的法律人格争议与传播侵权责任研究: 基于ChatGPT场景视角[J]. 南京社会科学, 2023(6): 90-99. [Google Scholar] [CrossRef
[4] 韩雨潇. 人工智能创作物版权侵权风险及其法律规制探析[J]. 科技传播, 2022, 14(13): 30-33. [Google Scholar] [CrossRef
[5] 周学峰. 生成式人工智能侵权责任探析[J/OL]. 比较法研究: 1-16. http://kns.cnki.net/kcms/detail/11.3171.d.20230714.1413.004.html, 2023-07-21.
[6] 王莹. 算法侵害责任框架刍议[J]. 中国法学, 2022(3): 165-184. [Google Scholar] [CrossRef
[7] 蔡雨恬. 论人工智能侵权中责任主体的认定与责任规制路径[J]. 中阿科技论坛(中英文), 2022(9): 195-200.
[8] 支振锋. 生成式人工智能大模型的信息内容治理[J]. 政法论坛, 2023(4): 42.