Banks and Lending Platforms Use AI Lie Detection to Combat Financial Black-Market Operations

Deep News19:21

Since mid-July, Wang Yu, a risk control officer at a regional branch of a rural commercial bank in East China, has been tasked with urgently integrating "smart online interview" technology into the existing consumer credit risk approval process. Working closely with technical colleagues, he has been evaluating the effectiveness of this application. For instance, after the risk control system asks a borrower, "Did the funds used to repay your overdue loans and restore your personal credit between December last year and March this year come from your salary or family support?" the AI is expected to capture changes in the applicant's facial expression upon answering "yes" to accurately determine if they are lying. If the AI deems the response deceptive, the consumer loan application will face stricter scrutiny and has a high probability of being rejected.

Over the past two weeks, Liu Wei, an operations director at a lending platform, has been busy negotiating with other platforms and licensed consumer finance companies to expand the sharing of delinquent borrower lists. This allows them to quickly identify if borrowers who have defaulted at other institutions are applying for loans at their own platform. "We call this 'data collision,' and it's essential for preventing loan default risks right now," Liu Wei said.

Since July, both Wang Yu's bank and Liu Wei's lending platform have noted an anomaly: some borrowers who had just restored their personal credit in the first quarter of this year applied for consumer loans between April and June, but then not only failed to meet their repayment obligations but also demanded substantial loan principal and interest reductions, citing reasons like "sudden illness making short-term work impossible." "It can't be that coincidental," Liu Wei remarked.

According to a notice from the People's Bank of China concerning the implementation of a one-time credit restoration policy, for loans overdue between January 1, 2020, and December 31, 2025, with a single overdue amount not exceeding 10,000 yuan, if the borrower fully repays the overdue debt before March 31, 2026, the record will no longer be displayed in the financial credit information database. What banks and lending platforms did not anticipate was that loan intermediaries would see an opportunity in this personal credit restoration measure. They provide high-interest private loans to help borrowers repair their credit, then "guide" these borrowers to apply for larger loans from formal financial institutions. Subsequently, the intermediaries earn substantial fees by helping borrowers drastically reduce the principal and interest on these loans through a "one-stop service."

Why just 10 ASX 200 shares?

Wang Yu received his urgent assignment because his branch's post-loan management department discovered that the default rate for first-time borrowers in June had surged more than six times the historical average. Previously, the monthly default rate for first-time borrowers was less than 0.5%—meaning fewer than five out of 1,000 new borrowers failed to repay. In June, that figure suddenly exceeded 3.5%, and the department noted that by July, it was likely to surpass 4%. After reporting this anomaly to branch leadership, the risk control team was brought in to investigate. Wang Yu's investigation revealed that the majority of first-time borrowers defaulting in June and July had one common characteristic: they had repaid their overdue loans and restored their credit in the first quarter of this year. Recently, many of these defaulting borrowers suddenly approached the bank, citing "sudden illness" to request reductions of over 50% on their loan principal and interest. "At that point, the post-loan department suspected that an invisible hand was orchestrating the entire process—from credit restoration, to applying for consumer loans of tens of thousands of yuan, to demanding drastic debt relief," Wang Yu said. However, the department lacked concrete evidence. The branch could only advise borrowers not to be misled by financial black-market organizations and ask the risk control department to strengthen "mid-loan risk management" to curb the escalating default risk.

Wang Yu learned from a risk control officer at a joint-stock bank's East China branch that the latter's branch had issued several million yuan in consumer loans to over 100 borrowers who had completed personal credit restoration in the second quarter. Since July, however, half of these borrowers had neither fulfilled their repayment obligations nor demanded principal and interest reductions, citing health issues. Liu Wei's lending platform faced a similar situation. Since April, the platform was delighted to see over 100 individuals with "no overdue credit records" proactively applying for loans via its mobile app. But since July, more than 15% of these borrowers had "disappeared" before their first repayment due date, ignoring all reminders. "Our collection department found it suspicious," Liu Wei said. After investigating, they discovered that most of the borrowers who defaulted in July had just repaid their overdue loans and restored their credit in the first quarter. As the lending platform analyzed this trend, lawyers from several law firms suddenly contacted them, claiming to represent borrowers and demanding substantial loan principal and interest reductions.

How to begin

At the beginning of this year, Chen Yong received a call from a loan advisory company offering a 10,000-yuan loan to help him repay an overdue amount of over 8,000 yuan from 2025, thereby restoring his personal credit. Chen Yong was tempted. For three years, he had worked as a barber with unstable income. In 2020, he took out an 8,000-yuan consumer loan from a bank to celebrate his girlfriend's birthday, but it became overdue. The bank agreed to waive penalty interest, but the principal had to be repaid by the end of 2026. When the barbershop he worked at suddenly closed last year, repaying the loan became even harder. After a friend's recommendation, he applied for an assembly worker position at a large electronics factory in the Jiangsu-Zhejiang region, offering a monthly salary of about 6,500 yuan. The company required a personal credit report, but his overdue record cost him the job. Desperate to fix his credit, Chen Yong signed a loan agreement with the advisory company in late January, borrowing 10,000 yuan at an annual interest rate of 36%. He used the money to repay the 8,000-yuan overdue bank loan. By the end of February, his credit was restored, and his report no longer showed overdue information. Feeling he could now find a better job, he searched for positions offering over 6,500 yuan per month but had no luck by March. Starting in April, the advisory company urged Chen Yong to apply for at least 30,000 yuan in consumer loans from banks or lending platforms. They even suggested a loan purpose: "skills training to improve job prospects." They advised him to apply to institutions where he had never borrowed before, as they focus on the credit records of first-time borrowers, making approval more likely if there are no overdue marks. Initially reluctant, Chen Yong was swayed when the company reminded him that their private loan agreement included a penalty interest rate of over 80% per annum for late repayment. To avoid a high-interest debt trap, he applied to three banks and lending platforms. Thanks to his clean credit record and a loan purpose of "purchasing training courses to enhance skills," he received credit lines of 15,000 yuan and 18,000 yuan from a bank and a lending platform, respectively. After receiving the funds in May, he quickly repaid all principal and interest to the advisory company. "There's no such thing as a free lunch," Chen Yong said. He borrowed 10,000 yuan for less than four months but ended up paying about 16,000 yuan, including penalties. Unexpectedly, in July, the advisory company contacted him again, proposing to "help significantly reduce the 33,000 yuan in loan principal and interest" in exchange for a 30% service fee on the reduction amount. The company even drafted a justification for him: "suffering a sudden illness, unable to work temporarily." When Chen Yong refused to fabricate an illness, the company told him that many borrowers in similar situations were using complaints to secure drastic debt relief.

Xiao Peng, operations director at a loan advisory company, has also targeted this "one-stop service" business. Starting last year, his company purchased large amounts of client data from lending platforms, screening for eligible borrowers. They offered private loans of up to 10,000 yuan at annual interest rates exceeding 30% to help them restore credit. Between December last year and March this year, Xiao Peng's company lent to about 150 borrowers. After their credit was restored, the company urged them to apply for tens of thousands of yuan in consumer loans from multiple banks and lending platforms, citing purposes like "skill learning" or "home renovation," but the real goal was to repay the private loan principal, interest, and penalty fees. When borrowers struggled to repay the new loans, the company offered to negotiate with banks and platforms for substantial debt relief, charging a service fee of about 40% of the reduction amount. Xiao Peng admitted that through this "one-stop service," the company makes money twice: once from the interest and penalties on the 10,000-yuan private loan, and again from the service fee. "For each deal, providing 10,000 yuan in principal, we collect over 10,000 yuan in total from interest, penalties, and service fees," he said. While his company's business is modest, others have earned over 6 million yuan in interest and penalties alone in the first half of the year. Xiao Peng noted that most loan advisory companies involved in this business have ties to financial black-market operations.

Is AI risk control effective?

Through days of evaluation, Wang Yu found that the smart online interview system has a high probability of "falsely accusing" honest individuals. For example, some borrowers either habitually pause for a few seconds before answering or look naturally nervous, prompting the AI to quickly conclude that the "borrower is lying." "In reality, our risk control department verified that these borrowers' answers were truthful," Wang Yu said. Additionally, many borrowers who restored their credit in the first quarter genuinely need new, larger loans for career transitions or household expenses. "We cannot treat all these individuals as being misled by financial black-market operations," he cautioned. In late July, Wang Yu suggested to branch leadership that for loan applications from credit-restored individuals, the bank should more thoroughly assess their actual living situation, focusing on future income, career development, repayment willingness, and capacity to make more scientific credit approval decisions. However, branch leadership raised concerns about how to accurately assess borrowers' future career and income. Instead of predicting the future, they preferred to focus on the present. Consequently, Wang Yu was tasked with communicating with other licensed consumer finance companies and banks to explore using privacy-preserving technologies to legally share delinquent customer lists, thereby quickly identifying if an applicant has recently defaulted elsewhere and avoiding risk.

Liu Wei, however, believes that sharing delinquent lists alone is a temporary fix. Over the past two weeks, he found that the lists provided by some institutions were incomplete, resulting in "slippage." His platform issued loans to several credit-restored borrowers only to later discover that they had defaulted at other platforms and were demanding debt relief for health reasons. "Strengthening mid-loan risk control alone is still not enough to prevent these issues," Liu Wei admitted. An effective solution would be to intensify efforts to crack down on financial black-market organizations at the pre-loan stage, significantly reducing their space for providing "one-stop services." Regulators are already taking action. On April 2, the Economic Crime Investigation Bureau of the Ministry of Public Security and the Inspection Bureau of the National Financial Regulatory Administration jointly held a video conference to launch a new wave of targeted crackdowns on financial "black and gray" industry crimes, further rectifying illegal financial intermediary chaos, firmly safeguarding financial order, and protecting consumers' legitimate rights and interests, aiming to secure high-quality financial development with high-level safety.

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Comments

We need your insight to fill this gap
Leave a comment