Lending Club AI Credit Scoring Used to Mask Bad Loans and Fake Borrowers

Who Gets Targeted

Primary victims include retail investors seeking returns higher than bank deposit rates, institutional financial advisors lacking professional risk management capabilities, and subprime borrowers attracted by low-threshold online loan applications. Retail investors often suffer from information asymmetry and a psychological tendency to blindly trust platform algorithm endorsements, assuming AI scoring can replace traditional manual bank reviews while ignoring the risks of underlying asset authenticity and insufficient bad debt provisions. Subprime borrowers, often in urgent need of liquidity or possessing weak credit histories, are misled by the platform's loose approval criteria, ultimately falling into a debt cycle of compounded high interest.

骗局怎么运作

• The platform publicly claims that its self-developed AI credit scoring system uses machine learning to accurately predict borrower default probabilities, thereby reducing bad debt rates and securing investor returns. In reality, the system incorporates large amounts of unstructured data and self-reported borrower information into the model, but the underlying asset review still relies on loose online processes, providing opportunities for fraudulent borrowers.
• Some internal employees or cooperative channel partners assist in forging borrower identity information, income statements, and employment records, exploiting the AI system's over-reliance on data completeness to bypass traditional face-to-face verification and manual due diligence. Externally, they emphasize 'fully online intelligent approval,' while internally, they lower manual review thresholds under the guise of 'optimizing model pass rates.'
• The platform packages high-risk loans into structured products for investors while simultaneously issuing new loans to cover maturing bad debts, creating a Ponzi-like capital pool. The external narrative is 'diversified investment to reduce risk,' while the reality is masking deteriorating underlying assets and liquidity gaps.
• When delinquency rates rise or regulators inquire, the platform exploits the explainability flaws of AI models, attributing bad debts to 'low-probability market fluctuations' or 'model transition periods,' and refuses to disclose underlying asset details or write-off data under the pretext of an 'algorithmic black box.'
• Once investors initiate mass redemptions or regulators trigger on-site inspections, the platform initiates emergency credit enhancement measures, such as introducing third-party guarantees or ABCP bridge financing. However, these funds often originate from related parties or new investors, further masking the true non-performing loan rate.
• Ultimately, after a concentrated outbreak of bad debts or internal whistleblowing, the platform is exposed for having a massive list of fake borrowers and provision gaps. Investor principal becomes unrecoverable, and management faces regulatory inquiries or criminal referral, completing the cycle of AI credit fraud.

红旗信号（看到这些快跑）

• 🚩 Lack of transparency in AI credit scoring model accuracy and bad debt data; only overall pass rates are disclosed without underlying asset write-off details.
• 🚩 Borrower qualification reviews are completed entirely online, lacking facial recognition, cross-verification with tax data, and manual face-to-face verification.
• 🚩 The platform frequently delays the release of delinquency data or modifies disclosure standards under the pretext of 'system upgrades' or 'model iterations.'
• 🚩 Claims of bad debt rates significantly lower than industry peers despite exorbitant loan interest rates, using high yields to cover hidden bad debts rather than genuine risk management capability.
• 🚩 Internal employees or channel partners can bypass approval by uploading fictitious payroll records and employment proofs, and the platform is not connected to central bank credit bureaus or third-party anti-fraud databases.
• 🚩 Investor redemption cycles are artificially extended or restricted, while the platform simultaneously launches new high-interest products to attract incremental capital.

真实案例

• In August 2026, the Economic Observer reported 'Abnormal Delinquencies,' noting that some AI credit platforms showed a divergence between delinquency rates and actual borrower income levels, suggesting the presence of mass fake borrowers or excessive model beautification.
• In June 2026, Sina Finance cited 32 regulatory guidelines, stating that an AI lending error by a city commercial bank led to a corporate capital chain rupture, exposing systemic risks in AI risk management lacking manual review and accountability mechanisms.
• In August 2026, Zhejiang police dismantled an AI online lending fraud gang, arresting 30 people involved in over 10 million yuan. The gang used AI technology and fake credentials to apply for loans in bulk before cashing out and disappearing, reflecting the severity of AI credit approval processes being targeted by illicit syndicates.
• In August 2022, the U.S. Federal Trade Commission (FTC) announced the return of over $9.7 million to 61,990 consumers charged hidden fees by LendingClub. The FTC sued the company in April 2018 for falsely promising 'no hidden fees' loans. This was the second distribution of compensation for the case, with total refunds exceeding $17.6 million. (Source: https://www.ftc.gov/news-events/news/press-releases/2022/08/federal-trade-commission-returns-more-97-million-consumers-harmed-lendingclubs-deceptive-hidden-fees)
• In July 2021, LendingClub reached a settlement with the FTC. The FTC accused the company of misleading loan applicants, concealing fee amounts, and unauthorized deductions. The settlement prohibited further false representations and required clear fee disclosure. The FTC initiated the claims process in February 2022 and distributed over $9.7 million to 61,990 consumers in August of the same year, with total refunds exceeding $17.6 million. (Source: https://www.ftc.gov/news-events/news/press-releases/2022/08/federal-trade-commission-returns-more-97-million-consumers-harmed-lendingclubs-deceptive-hidden-fees)

Official Stance

• In March 2026, the China Banking and Insurance Regulatory Commission (NFRA) interviewed five major loan facilitation platforms including Lexin, pointing out issues with disguised high interest rates and opaque algorithms, emphasizing that AI risk models must be explainable and traceable.
• In June 2026, regulators issued 32 guidelines on AI lending liability for banks, clarifying the responsible parties for AI approval errors and provision requirements, prohibiting platforms from evading bad debt liability under the guise of an algorithmic black box.
• In August 2026, Xinhua News Agency's Economic Information Daily issued a warning against financial illicit syndicates targeting 'credit-blank' individuals, exposing the chain of fake loan applications and cashing out, urging financial institutions to strengthen anti-fraud verification in AI approval processes.

How to Protect Yourself

• ✅ Investors should demand that platforms disclose underlying asset details, write-off data, and AI model accuracy audit reports, refusing to make investment decisions based solely on yield marketing.
• ✅ Borrowers should apply for loans through licensed financial institutions or channels connected to the central bank's credit bureau, avoiding the submission of ID and bank card information to platforms without lending qualifications.
• ✅ Regulators could require AI credit platforms to submit periodic reports on bad debt and fake borrower ratios audited by independent third parties, and make the audit results public.
• ✅ Financial institutions should set manual review red lines for AI approval results, mandating face-to-face verification and tax cross-verification for high-amount or high-frequency loan applications.
• ✅ The public should regularly check their personal credit reports and immediately report to the central bank's credit center and public security organs if they discover their identity has been used for fraudulent loan applications.