人工智能审计系统建设思路及实践探索  

The Ideas and Practical Exploration of AI Smart Auditing Construction System

作　　者：黄小忠 吕全 刘军 曾雪峰 吕正东 张正锋 梁若冰 Huang Xiaozhong;Lyu Quan

机构地区：[1]广西壮族自治区审计厅,530022 [2]杭州数令集科技有限公司,311122

出　　处：《审计研究》2024年第6期30-38,共9页Auditing Research

摘　　要：随着人工智能技术的迅猛发展,特别是在大模型领域的突破,新技术的应用已经渗透到诸多行业和各个领域,如何应用人工智能技术赋能审计工作,促进审计事业高质量发展,提高审计工作效率、准确性和决策质量,成为当前的重要课题。本文旨在探讨和介绍如何构建以大模型技术为核心的人工智能审计系统,运用审计大模型在理解、对话、推理和生成方面的能力,对审前准备工作、现场审计作业、审计项目审理以及审计项目后评估等审计工作关键流程进行智能化辅助,并通过广西壮族自治区审计厅的实践案例,展示系统建设过程中的思路、方法和创新点,为人工智能技术在审计领域的落地应用提供可借鉴的方法。With the rapid development of Artificial Intelligence(AI) technology, especially breakthroughs in the field of large language models, the application of new technologies has penetrated various fields of many industries. It has become an important issue as to how to apply AI technology to empower auditing work,promote high-quality development of the auditing profession, and improve the efficiency, accuracy, and decisionmaking quality of auditing work. This paper aims to explore and introduce how to build an AI smart auditing system with large language models technology at its core. By applying the capabilities of auditing large language models in understanding, dialogue, reasoning, and generation, it provides intelligent assistance in key processes of auditing work such as pre-audit preparation, on-site audit operations, audit project quality review, and postaudit evaluation. Through practical cases from the Audit Office of Guangxi, it showcases the thought process,methods, and innovations in the system construction process, offering a reference for the practical application of AI technology in the auditing field.

关 键 词：人工智能审计系统 大模型 审计专家系统 融合框架 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] F239.1[自动化与计算机技术—控制科学与工程]