On the afternoon of September 15, the first literature study on future rule of law and digital intelligence technology innovation was successfully held in the conference room of Legal Big Data and Artificial Intelligence Laboratory of Southeast University School of Law. The theme of this reading activity is “Can AI be Audited”, and the work of literature reading was undertaken by Associate Professor Yang Jie, and Cui Rongzhen, a 24th grade doctoral student of the School of Law. Prof. Wang Lusheng, Director of Social Sciences Office, Associate Prof. Xu Minchuan, Ms. Bi Sheng and students from Digital Law Team participated in this activity.

Ms. Yang Jie firstly introduced the core content of the research literature, compared the concepts, contents and functions of AI auditing in the primary and secondary literature, and sorted out the more complicated algorithmic auditing system in practice with a categorized perspective; at the same time, she pointed out that the accuracy of the audit report is yet to be examined due to the lack of auditor's responsibility system in reality. Ms. Yang interacted with students on the social risks derived from AI auditing.

Cui Yongzhen introduced in detail the line of thought and the main content of the literature, and shared the experience of the study. Cui pointed out that the main reading literature's introduction to the functions and norms of AI auditing reflects a high degree of knowledge and synthesis characteristics, and also constructs a countermeasure framework for the inadequacy of the AI auditing system. The main reading literature has made a detailed compilation of the institutional history of financial auditing, as well as the AI auditing legislation and typical cases in major Anglo-Saxon countries, which provides readers with rich theoretical and practical materials.

In the exchange and discussion session, teachers and students exchanged warmly and the academic atmosphere was strong.Zhou Yuheng, a PhD student of grade 23, expressed his insights on the subjects of AI audit and analyzed the roles played by different subjects in AI audit according to their respective interests and needs.Wang Jingyi, a PhD student of grade 24, believed that AI audit is an extension and attempt of the general theory of audit in the theory of algorithmic governance, and that the coerciveness and authoritativeness of the audit can somewhat degree can balance the national arithmetic and enterprise arithmetic, and introduces the legislative development of AI auditing in comparative law.24 Xiao Yuanhang, a PhD candidate of class 24, analyzes AI auditing through three levels of sub-scenarios, sub-layers, and sub-targets.22 Sun Yu, a PhD candidate of class 22, believes that the concept of AI auditing is deficient in theoretically proposing justification, and that it can be further explored in terms of its theoretical support, and digging into the ways to reconcile its general function with specific decision-making.Wang Yang, a Ph.D. student in class 23, believes that algorithmic auditing may help in the criminal justice field by classifying and configuring different risks, and provides ideas for domestic related research. The rest of the students had an intense discussion around the topic of AI auditing.

Mr. Bisheng highly affirmed the logic and writing specification of the main literature, and pointed out that the auditing approach may cause security guardrails, which may affect the performance of AI models. He then explained the “algorithmic black box” problem from the technical level, arguing that the technical feasibility of auditing large models mentioned in the literature is not high, and it can only be regulated by using additional means such as post-processing; he also supported the view that the public power acts as the auditing subject, and put forward further thoughts on the cycle of AI auditing.

Mr. Xu Minchuan firstly pointed out that the way of reading club can drive teachers and students to read academic literature, strengthen academic taste, enhance academic ability, and ultimately systematically improve the ability of academic output, and encourage everyone to better participate in the literature reading activities in the future. Secondly, as far as this literature is concerned, the audit system has a more mature supervision and operation mechanism, AI audit can be endorsed by the national public authority of the auditing unit, so as to play the legal effect of the legal system.

Mr. Lusheng Wang emphasized that the rise of big models nowadays makes the discussion on the topic of AI auditing more contemporary. Modernity reduces people's judgment of complexity in various ways, and auditing is undoubtedly one of these simplifying mechanisms. The understanding of science and technology is mainly divided into three major schools of thought: autonomy theory, instrumental theory, and relational theory, with different views fundamentally affecting the legal constraints on science and technology, and AI auditing is essentially the regulation of science and technology by human beings, which is the classic proposition of “law and science and technology” in jurisprudence. Afterwards, he affirmed and analyzed Mr. Zhang Tao's article, and thought that his writing style and the theoretical dialogues in the article were worth learning together, and shared the methods of academic innovation with everyone.

At the end of the activity, Ms. Yang Jie summarized the literature reading activity. Organizing the literature reading activity, on the one hand, makes everyone have a common topic for discussion and promotes the interaction between teachers and students; on the other hand, it can also be more inspiring for teachers to write their articles. The first session of the future rule of law and digital intelligence technology innovation literature reading activity was successfully concluded in the warm discussion between teachers and students.