人工智能模型栈在国家审计中的应用  

The Application of Artificial Intelligence Model Stack in National Auditing

作　　者：朱琢 Zhu Zhuo

机构地区：[1]江苏省无锡市审计局,214000

出　　处：《审计研究》2025年第1期51-56,共6页Auditing Research

摘　　要：近年来,人工智能技术和应用快速发展,ChatGPT等应用不断突破。经历了传统财务审计、计算机辅助审计到大数据审计的审计技术跃迁,无锡市审计局在养老、医保等民生审计项目中,不断寻求人工智能与审计的交汇点。通过多维分析审计对于人工智能的功能需求,在审计重点研判、疑点数据分析、文字图像处理等方面探索人工智能技术和工具在项目实施中的应用,提出了人工智能应用功能模型栈理念。不同于常见的一体化人工智能应用产品,提出依托分层低耦合架构,让各类人工智能技术应用融合审计项目全流程,根据项目需求灵活选用生成式人工智能、自然语言处理、机器学习、深度神经网络等技术的审计实践,有效提升审计数据分析效率,减轻了审计人员负担。In recent years,artificial intelligence(Al)technology and applications have developed rapidly,with breakthroughs in applications such as ChatGPT.The audit technology has evolved from traditional financial auditing,computer-assisted auditing,to big data auditing.The Wuxi City Audit Bureau has been continuously seeking the intersection of Al and auditing in audit projects related to the people's wellbeing such as elderly care and medical insurance.By conducting multidimensional analysis of the functional requirements of AI in auditing,and exploring the application of AI technology and tools in project implementation in areas such as key audit judgment,suspicious data analysis,and text image processing,the concept of an AI application functional model stack has been proposed.Unlike common integrated AI application products,a layered,loosely coupled architecture is proposed to integrate various AI technologies into the entire process of audit projects.According to project needs,auditors can flexibly select and apply technologies such as generative Al,natural language processing,machine learning,and deep neural networks,to effectively improving the eficiency of audits related to the people's wellbeing,especially in audit data analysis,and thus reducing the burden on auditors.

关 键 词：人工智能应用 功能模型栈 民生审计 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] F239.44[自动化与计算机技术—控制科学与工程]