A credit evaluation model is used to automate and quantify a personal credit evaluation, and the results can have a significant impact on a person's overall economic and financial situation. In order to prevent financial customers from suffering as a ...
A credit evaluation model is used to automate and quantify a personal credit evaluation, and the results can have a significant impact on a person's overall economic and financial situation. In order to prevent financial customers from suffering as a result of inaccurate credit evaluation results, it is important to ensure the fairness and dependability of the credit evaluation model.
The recently developed AI methodology-based credit evaluation model, in contrast to the existing credit evaluation model, has the advantage of reflecting a variety of non-financial information and performing relatively well, but it also raises questions about the model's bias and is challenging to interpret and explain. Therefore, it is necessary to examine new issues that may arise from the perspective of financial consumer protection.
In this study, domestic and foreign laws and systems were compared and analyzed from the perspective of financial consumer protection to confirm that the main principles are the prohibition of discrimination, the right to request an explanation, and the right to request a correction. In particular, it was confirmed that the main issues are the overfitting risk and bias due to the characteristics of the artificial intelligence methodology and determining the level of algorithm transparency to improve explanatory power. On the basis of this, regulatory approaches were suggested in order to reflect new financial consumer protection issues that may occur as a result of the advancement of personal credit evaluation models in the domestic discipline system.