生成式人工智能在征信领域的应用研究:前景、风险与对策  

Research on the Application of Generative Artificial Intelligence in the Field of Credit Reporting:Prospects,Risks and Countermeasure

作　　者：丁略涛 Ding Luetao(Guanghua Law School,Zhejiang University,Hangzhou 310008,Zhejiang,China)

机构地区：[1]浙江大学光华法学院,浙江杭州310008

出　　处：《征信》2024年第8期27-36,共10页Credit Reference

基　　金：国家社会科学基金重大项目(21&ZD199);国家重点研发计划项目(2021YFC3340300)。

摘　　要：生成式人工智能凭借强大的自主学习能力、关联分析能力、结果输出能力,在征信领域具有提高信息输入效率、提升信用评估准确性、生成与解读信用报告、预知风险的应用潜能,但也面临着多重挑战。结合技术原理与征信领域的特性,可从输出结果的质量问题、数据泄露的安全问题、信息处理的合规问题三个方面剖析征信领域应用生成式人工智能的风险隐患。为平衡创新发展与风险防范,应在规范体系层面倡导目标导向式的框架性立法、以人为本确立基本原则,在治理主体层面倡导协同共治、多元参与,在监督管理层面倡导包容审慎的监管立场与分类分级的监管方式并行。With strong self-learning ability,correlation analysis ability and result output ability,generative artificial intelligence has the application potential of improving information input efficiency,improving the accuracy of credit assessment,generating and interpreting credit reports,and predicting risks in the field of credit reporting,but it also faces multiple challenges.Combined with the technical principle and the characteristics of the credit reporting field,the potential risks and vulnerabilities of applying generative artificial intelligence can be analyzed from three aspects:the quality of the output results,the security of data leakage and the compliance of information processing.In order to balance innovative development and risk prevention,we should advocate goal-oriented framework legislation at the level of the normative system,establish basic principles based on people,advocate collaborative governance and multiple participation at the level of the governance body,and advocate inclusive and prudent regulatory stance and classification and classification supervision at the level of supervision and management.

关 键 词：生成式人工智能 征信 创新发展 风险预防 合作治理 

分 类 号：F832.4[经济管理—金融学] D912.1[政治法律—宪法学与行政法学]