首页|生成式人工智能(AIGC)下的个人信息保护行政公益诉讼探索与发展

生成式人工智能(AIGC)下的个人信息保护行政公益诉讼探索与发展

扫码查看
伴随着生成式人工智能的崛起,无论是训练数据集的跨域、多模态,还是大模型应用潜在的算法歧视,都给传统个人信息保护模式带来了多维度的挑战,极大地限制了告知同意、目的限制等原则的现实价值.个人信息保护虽被纳入检察公益诉讼管辖范畴,但生成式人工智能的出现导致主体责任分散多元化等问题,需要行政公益诉讼制度作出回应.基于对大模型"幻觉"与多模态、跨域数据集的分析,个人信息保护行政公益诉讼应在预防性规制和协作机制方面进行完善.具体的制度完善建议主要集中在建立跨行政区划检察案件管辖体系、"统-分"式行政公益诉讼监督网、类案监督治理模式以及以公共安全为主体的事前预防模式.
Exploration and Development of Administrative Public Interest Litigation for Personal Information Protection under AIGC
With the rise of AIGC,whether it is the cross-domain and multi-modal training data sets,or the po-tential algorithm discrimination in the application of large models,it has brought multidimensional challenges to the traditional personal information protection mode,which greatly restricts the practical value of the principles of consent and purpose limitation.Although personal information protection has been included in the jurisdiction of procuratorial public interest litigation,the emergence of AIGC has led to problems such as the dispersion and di-versification of subject responsibilities,which requires the administrative public interest litigation system to re-spond.Based on the analysis of the"hallucination"in large models and multimodal cross-domain datasets,the ad-ministrative public interest litigation for personal information protection should be perfected in terms of preventive regulation and coordination mechanism.The specific system improvement suggestions mainly focus on the cross-administrative division procuratorial case jurisdiction system,the"unified and divided"administrative public in-terest litigation supervision network,the case-type supervision management mode,and the pre-prevention mode with public safety as the main body.

AIGCpersonal information protectionpublic interest litigationlarge model

唐林

展开 >

合肥工业大学 文法学院,合肥 230011

生成式人工智能 个人信息保护 公益诉讼 大模型

2024

安徽理工大学学报(社会科学版)
安徽理工大学

安徽理工大学学报(社会科学版)

CHSSCD
影响因子:0.233
ISSN:1672-1101
年,卷(期):2024.26(6)