Measuring ethics level of technological topics using phylogenetic tree

Abstract

ABSTRACT This study attempted to construct a finance AI technology phylogenetic tree that represents the evolutionary pattern of finance AI technology and quantitatively measures the level of ethical attributes in the tree’s technological topics. Based on the tree, it was confirmed that the finance AI technology field is evolving in the direction of three research topics in 2021: deep learning, machine learning, and decision-making technology. Through NLP methodology, we proposed a model to measure the extent to which fairness, human right and normality, transparency and accountability, privacy and security, explainability, and performance guarantees are discussed among these topics. Results confirmed that explainability and performance guarantees were mostly discussed in the case of deep learning technology, privacy and security in the case of machine learning technology, and human rights and normality in the case of decision-making technology. Additionally, it was confirmed that the discussion on fairness in the overall finance AI technology field was low, suggesting that this discussion should be supplemented. The model presented in this study makes an important contribution in that it is capable of detecting the evolutionary direction of AI technology and quantitatively analysing their ethical attributes.