Ethical Dilemmas and Governance Paths of Algorithms in Fintech Innovation

Abstract

With the increase in the efficiency of services in recent years, many ethical problems have begun to appear in the algorithm system of fintech, such as a lack of transparency, unclear division of responsibilities and price discrimination; prediction inaccuracies have occurred due to sample bias, liquidity suction in high-frequency trading and erosion of privacy boundaries caused by cross-domain data fusion. Based on the analysis of the generation mechanism and transmission logic for the problems mentioned above, this paper puts forward three levels of governance: organising the development process with integrated ethical constraints, an adaptive mechanism for a dynamic interpretability framework, and an algorithm audit and responsibility allocation model based on a reputation system. The above paths form a closed loop from start to finish and offer an organised way to deal with the ethical problems of algorithms.