CLPS Completes AI-Assisted AML Review Project for Major Bank Using Fine-Tuned LLM
CLPS Incorporation (Nasdaq: CLPS) has completed an AI-assisted Anti-Money Laundering (AML) project for a major Chinese commercial bank, using fine-tuned small language models to achieve over 90% accuracy in risk ratings. The initiative was conducted within Hong Kong's Gen A.I. Sandbox and aimed at addressing the growing complexity of financial transactions while adhering to strict regulatory requirements on transparency and explainability. The project tackled challenges such as limited GPU resources, subjective evaluation standards, and low data availability by leveraging open-source small LLMs with domain-specific fine-tuning. CLPS also used synthetic data augmentation techniques like risk distribution matching and generative rewriting strategies to enhance model performance without compromising accuracy or traceability. By decoupling tasks and optimizing reasoning chains, the team reduced common AI issues such as hallucinations and context interference. The success of this project highlights a new fintech paradigm where compact models with robust fine-tuning outperform larger LLMs in cost-effective AML compliance solutions.
Tickers: $CLPS
Topics: AI deployment AML compliance banking innovation Generative AI