首页|生成式人工智能的数据治理与"囚笼"——基于行业主体自我监管的回应型研究

生成式人工智能的数据治理与"囚笼"——基于行业主体自我监管的回应型研究

扫码查看
以ChatGPT为代表的生成式人工智能自面世伊始即引人瞩目,数据、算法、算力作为推动生成式人工智能前行的"三驾马车"对其重要性不言而喻.然而,生成式人工智能引发的数据之殇,如数据泄露、数据垄断等亦挑战着传统数据治理框架.在智能时代,基于应用场景的分散治理、基于风险预防的事前治理、基于公权主导的硬性治理均难以满足治理与发展的双重需要.行业主体基于其专业性、及时性、协同性等灵活优势理应在智能时代的数据治理中发挥积极作用,却饱受自监自管、缺乏动力等诟病.目前,新加坡的问责制协同治理模式、美国的软法治理模式均反映着行业主体自我监管的可行性,且证实了自我监管并非完全自治.因此,以自我监管为着力点,从微观、中观、宏观三个角度构建数据治理新范式有其现实意义.
Data Governance and the "Cage" for ChatGPT-like Generative Artificial Intelligence:Responsive Research Based on Self-Regulation by Industry Subjects
The generative artificial intelligence represented by ChatGPT has been eye-catching since its inception,and the importance of data,algorithms,computing power as a the"troika"to promote generative artificial intelligence is self-evident.However,the death of data triggered by generative AI,such as data leakage and data monopolization,has also challenged the traditional data governance framework.In the era of intelligence,decentralized governance based on application scenarios,ex ante governance based on risk prevention,and rigid governance based on public authority dominance can hardly meet the dual needs of governance and development.Industry subjects should play an active role in data governance in the smart era based on their flexible advantages of professionalism,timeliness,and synergy,but they have been criticized for self-monitoring and self-regulation and lack of motivation.At present,Singapore's collaborative governance model of accountability and the U.S.soft law governance model reflect the feasibility of self-regulation of industry subjects and confirm that self-regulation is not completely autonomous.Therefore,it is of practical significance to construct a new paradigm of data governance from micro,meso and macro perspectives with self-regulation as the focus point.

ChatGPTgenerative artificial intelligencedata governanceself-regulation

胡裕岭、姚浩亮

展开 >

华东政法大学刑事法学院,上海 201620

ChatGPT 生成式人工智能 数据治理 自我监管

国家社会科学基金重大项目

20&ZD199

2024

征信
中国人民银行郑州培训学院

征信

CHSSCD北大核心
影响因子:0.904
ISSN:1674-747X
年,卷(期):2024.42(2)
  • 50