Today, my topic is “Exploring the Path to Challenging Fund-Related Expert Opinions Through the Review Mindsets of Accounting Forensics and Electronic Data Forensics.” I’ve been practicing criminal defense for over a decade, and in recent years, I’ve also been studying Professor Liu Pinxin’s approach to handling cases involving electronic data. Over time, I’ve accumulated some valuable insights, and today I’d like to take this opportunity to share them with all of you.

I will elaborate from three aspects, covering a total of nine sub-points.

First, I believe that the fund analysis and appraisal opinion (including the appraisal report and inspection report) possesses characteristics of both accounting appraisal and electronic data appraisal. On this basis, The necessity of having a fund analysis and appraisal opinion is mainly reflected in the following three aspects.

First, the computing power of traditional accounting verification is already insufficient. Today, we live in an era of massive electronic data, and the traditional manual methods of accounting calculations and audits can no longer keep pace with the times. Moreover, the findings and conclusions reached by accounting audits often remain confined to superficial aspects such as monetary amounts, transaction flows, and the identities of those involved. However, to look beyond these surface appearances and get to the heart of the matter—tracing the ultimate destination and ownership of funds—is something that accounting expertise alone simply cannot achieve.

Second, funding data has the form of electronic evidence and specific requirements for evidence collection. Financial data mostly exist in electronic form—such as bank statements and transfer records—all of which are presented as electronic data. The collection of electronic data comes with specific requirements; for example, the “Regulations on Several Issues Concerning the Collection, Extraction, and Examination and Adjudication of Electronic Data in Criminal Cases,” issued jointly by the Supreme People’s Court, the Supreme People’s Procuratorate, and the Ministry of Public Security. Failure to comply with these relevant regulations could directly affect the authenticity of the data as well as the validity of related evidence and expert opinions. Precisely because of these unique characteristics, it is necessary to establish electronic data examination as a distinct category of forensic expertise.

Third, the need for evidence conversion by public security organs. Fund analysis and identification can reveal interpersonal relationships by tracing the flow of funds. Take pyramid scheme cases as an example: the flow of funds in a pyramid scheme—through whom to whom, even passing through several layers—closely correlates with the hierarchical structure and the number of participants involved. The ultimate goal of analyzing fund flows and personnel relationships is not merely to determine the number of individuals or the amount of money involved; more importantly, it’s to identify the underlying behavioral patterns and the specific individuals implicated in the case. When public security authorities initiate an investigation, they may initially have only one or two suspects in hand. However, by using the method of tracing fund flows and uncovering the associated personnel relationships, all individuals involved in the case can be brought to light and brought under prosecution. If this behind-the-scenes work remains hidden and unexposed, it cannot serve as a basis for prosecuting other individuals involved in the case.

1. Fund analysis is a means, not an end.

Fund analysis is merely a tool. As mentioned above, its fundamental purpose is to identify and verify the behavioral patterns of criminal activities, the individuals involved, and the relationships among these individuals.

2. Proactive role of expert opinions
 

In the past, traditional forensic assessments typically played a role only at the later stages—for example, after the public security authorities had conducted a certain level of investigation and needed to determine such matters as the nature of injuries and the extent of damages—only then would they carry out forensic examinations to evaluate the consequences of the criminal act. Nowadays, however, financial analysis and forensic examinations are increasingly being conducted earlier in the investigative process, even during the preliminary investigation stage before formal case filing. These examinations serve to provide investigators with leads and assist them in their inquiries. Moreover, the forensic examination process and the investigative process may even overlap and proceed concurrently.

3. The fund analysis and identification process is dynamic.
 

Fund analysis and identification is an ongoing process of trial and error and verification. Take pyramid scheme cases as an example: In the past, accounting appraisals might have focused solely on identifying amounts of money involved and the number of individuals affected. But today, fund analysis and identification has become a tool to assist investigative efforts. Once funds are traced to specific individuals, public security authorities will investigate those individuals to determine whether they are suspected of involvement in the case. If no suspicion arises, that line of inquiry is abandoned; if suspicion does emerge, the investigation continues along that particular track. The "Regulations on the Procedures for Fund Analysis and Identification by Public Security Authorities (Implementation)" explicitly stipulate that any appraisal opinion used as evidence must be disclosed to the parties concerned. However, not all fund analysis and identification opinions will ultimately be admitted as evidence—only those deemed valid will serve as evidence, while invalid ones will not.

Since the types of forensic examinations are currently limited to four categories—forensic medical examinations, physical evidence examinations, audio-visual data examinations, and environmental damage assessments—it is not yet entirely clear whether financial analysis examinations should be classified as forensic or industry-specific examinations. Therefore, they may not fully comply with the General Provisions on Forensic Examination Procedures. However, according to the Regulations, financial analysis examinations share the same requirements as the General Provisions in terms of commissioning procedures, sample material requirements, signatures of examiners, and institutional seals. In addition to drawing on the review approaches used for existing forensic examination categories, given that financial analysis examination opinions combine features of both accounting examinations and electronic data examinations, we might explore potential avenues for challenging such opinions by adopting the review methodologies typically applied to accounting and electronic data examinations.

1. From the perspective of the review mindset in accounting forensic examination, does the fund analysis method possess scientific validity?

Scientific rigor encompasses both normativity and verifiability. In terms of normativity, although court rules clearly outline procedures for extracting, cleansing, and verifying financial data, the expert opinions presented merely describe the written workflows. As a result, we lawyers are unable to conduct a substantive review of the testing methods based on these documents—and even determining how to carry out such a substantive review is beyond our capabilities. At present, when conducting financial analysis on massive datasets, human resources are limited. In addition to leveraging big data, we also rely on algorithms and models. However, terms like “algorithm” and “model” tend to make us rather sensitive, given that in recent years, AI software and products have proliferated rapidly, sparking considerable controversy and criticism. Take AI algorithms as an example: the underlying algorithmic principles are often highly complex, and even the technical personnel who design these algorithms may not be able to clearly explain the specific mechanisms at work during actual operation. Moreover, when AI arrives at a particular conclusion, its scientific validity and rationality cannot be guaranteed with absolute certainty. For instance, in one case, the expert set a parameter of 0.2 when analyzing statistical data and used this value to screen other data. When asked about the source of this parameter and the method used to calculate it, the expert replied that he had no practical experience and simply felt that 0.2 was a reasonably sound choice. From this, it’s evident that the entire process resembles a series of hypotheses and inferences. Yet, its reliability and ability to withstand scrutiny and verification remain highly questionable.

2. Examining the Issue of Data Sources from the Perspective of Reviewing Expert Opinions on Electronic Data Authentication

The “Regulations” stipulate that public security authorities may conduct financial analysis and data authentication. When necessary, the Supervisory Commission, the Procuratorate, and the courts may also entrust such authentication services. The prerequisite for this entrusted authentication is that the Procuratorate, the courts, or the Supervisory Commission have already obtained the electronic data relevant to the case; only then can they commission the authentication to the internal forensic institutions of the public security authorities—this practice is entirely acceptable. There’s a practical situation worth noting: According to available information, the Ministry of Public Security has established a “Platform for Investigating and Controlling Funds Involved in Illegal and Criminal Activities.” This platform is linked to data from various financial institutions, greatly facilitating the work of public security agencies in case handling and investigation. For example, when a sum of money enters the bank account under Zhang San’s name, accounting verification typically stops there, concluding that Zhang San is the recipient of the funds. However, by analyzing data through the big-data platform, it’s possible to look beyond this surface-level observation and uncover that this account subsequently engaged in multiple fund transfers with an account registered under Li Si’s real name. Once such connections form a stable pattern, it becomes compelling evidence to infer that Zhang San’s account is actually controlled by Li Si—and even that funds originating from upstream sources are ultimately pooled into Li Si’s account. This, then, highlights the significance of big-data analysis of fund transfer records: It no longer confines itself to the relationship between the recipient of funds and the funds themselves, but rather enables us to penetrate beyond individuals to reveal the underlying relationships behind them.

This clearly demonstrates the importance of the platform for tracing and controlling funds involved in illegal and criminal activities: on the one hand, it can clearly pinpoint the ultimate destination of these funds; on the other hand, when public security authorities are handling a case, they can seamlessly upload the data they have collected onto this platform for comparison, thereby facilitating both investigation and evidence gathering. Extending from this issue, if the Supervisory Commission, the People's Procuratorate, or the People's Court, when commissioning forensic examinations, need to rely on this “platform for tracing and controlling funds involved in illegal and criminal activities,” how should the procedural links be coordinated? And how can the issue of the legality of data sources be addressed?

3. Gain a moderate understanding of the mindset for handling cases involving electronic evidence, and leverage the expertise of expert witnesses.

As previously mentioned, artificial intelligence also faces numerous challenges, including algorithmic bias, hallucinations, and poor transparency—issues that go beyond the scope of a lawyer’s expertise. What lawyers can typically do is seek assistance from experts. However, finding such experts often presents a host of difficulties.

First is the issue of expert qualification certification. Currently, experts in society do not possess the qualifications required to serve as forensic data analysts, and the standards for expert witnesses recognized by the courts remain unclear. Several years ago, when handling a case involving the infringement of citizens’ personal information, I sought out an expert who had previously served as a communications security officer at the G20 summit and had received numerous national honors, intending to call him as a witness. I submitted this expert’s certificates of honor to the court, but they were not accepted. The court stated that experts must hold qualifications as forensic analysts. However, the law does not explicitly stipulate that experts must have such qualifications; rather, this requirement largely reflects the individual judgment of judges. This requirement could potentially hinder our ability to secure suitable expert witnesses.

Next is the issue of selecting the type of expert. Although lawyers may not be technically proficient, they need to understand the mindset behind handling cases involving technology or electronic evidence. They must clearly identify which matters require expert assistance and in which areas experts can provide expert opinions, verification, and technical processing. Only then can they precisely match the right experts to the case. Otherwise, for instance, if they’re unclear about even fundamental issues—such as whether ink from a seal can be dated or how it differs from chronological analysis of red and black ink—they’ll find it extremely difficult to proceed with their work.

Finally, there’s the question of how experts can play their part. Take the retrieval of evidence as an example: The law does not explicitly stipulate lawyers’ rights regarding the retrieval of evidence, yet evidence is indispensable for forming expert opinions. Based on my successful experience in retrieving evidence during case handling, lawyers can only secure the judge’s approval to retrieve evidence if they raise sufficient doubts about the expert opinion—doubts that are compelling enough to make the judge realize that although retrieving evidence may increase the judge’s workload, it could potentially influence the conclusions of the expert opinion. If lawyers fail to shake the judge’s confidence, obtaining the evidence becomes impossible, and even experts will find it difficult to play a meaningful role.