生成式人工智能应用中个人信息保护面临的挑战与应对
The Challenges and Countermeasures of Personal Information Protection in the Application of Generative Artificial Intelligence
马恩霖1
作者信息
- 1. 海南大学 法学院,海南 海口 570228
- 折叠
摘要
在ChatGPT发布的推动下,全球人工智能产业迎来了生成式人工智能应用研发的繁荣期.当前,提升生成式人工智能能力的核心在于数据的收集与利用.然而,从机器学习到用户交互,生成式人工智能应用的数据处理全过程对个人信息保护构成了严峻挑战:知情同意规则的虚化、目的限制原则的忽视、公开透明原则的架空,以及过度数据挖掘行为与最小必要原则的冲突.本文认为应当建立公开信息的合理使用框架,推行去标识化训练数据制度,采用风险分级治理的监管策略以实现各等级生成式人工智能应用的差异化监管,并强化算法的可解释性,以应对生成式人工智能应用发展中对个人信息保护的挑战.
Abstract
The launch of ChatGPT ushered in a boom in the development of generative AI applications in the global AI industry,a technological innovation hailed as the core driving force of the fourth technological revolution.At present,the core of improving the ability of generative artificial intelligence lies in the collection and efficient use of data.However,from machine learning to user interaction,the whole process of the data processing of the generated AI reference poses a serious challenge to the protection of personal information:the weakening of the rules of informed consent,the neglect of the principle of purpose restric-tion,and the absence of the principle of openness and transparency,and the conflict between excessive data mining behavior and minimum necessary principle.It is suggested that a framework for the rational use of open information should be established,a de-labeling training data system should be implemented,and a risk-graded governance strategy should be adopted to achieve differentiated supervision of AI applications at all levels,the interpretability of the algorithm is enhanced to meet the challenge of personal information pro-tection in the development of generative artificial intelligence applications.
关键词
生成式人工智能应用/个人信息保护/知情同意规则Key words
Application of Generative Artificial Intelligence/Protection of Personal Information/The In-formed Consent Rule引用本文复制引用
出版年
2024
NETL
NSTL
万方数据