中圖分類號： F239
文獻標識碼： A
DOI： 10.19358/j.issn.2097-1788.2023.02.008
引用格式： 付冉冉. 大數據時代算法審計構想[J].網絡安全與數據治理，2023，42(2)：48-52.

The conception of algorithm audit in big data era

Abstract： In the modern context, algorithm is an automatic program to solve problems, which is widely used in public management, judicial adjudication, financial system, business and so on. On the one hand, the application of the algorithm improves the operation efficiency of the whole society;on the other hand, it also leads to the disorder of social management because of its complex technical characteristics. At present, the legislation of the algorithm regulation system in our country is relatively loose, and the supervision subject is not clear enough, and audit, as an independent and mature examination system, has great potential in the supervision of algorithm design, algorithm data protection and legal accountability of algorithm decision-making. It can solve the problems such as loose algorithm supervision system and unclear supervision responsibility. Under the audit supervision mode, professionals can formulate audit plans, conduct risk assessment and factual review according to specific audit standards, give audit conclusions, produce audit reports, conduct audit publicity, and implement accountability to relevant personnel and follow-up audit quality follow-up. By the combination of third-party audit, platform self-regulation and industry self-discipline, algorithm supervision can usher in a new breakthrough.

Key words : algorithm；algorithm risk；algorithm governance；algorithm audit