A research on algorithmic bias and fairness in machine learning-based support decision systems: A case analysis of credit evaluation

In this study, the increasing exposure to artificial intelligence, machine learning, and algorithms that individuals commonly encounter in their daily lives is addressed. However, with the proliferation of the use of these technologies, new concepts such as algorithmic bias and machine learning unfairness have emerged. While the decisions taken by machine learning in a limited area can make a difference, it is stated that with the expansion of its scope, the impact of algorithms can go beyond narrow domains and lead to societal divisions. Therefore, machine learning algorithms used in credit decision-making processes were tested for bias or inequality on German credit data. Test results were interpreted using statistical results and inequality metrics, and it was determined that machine learning algorithms cause unfairness in credit processes. It has been discussed that algorithmic biases may arise from the lack of extensive learning and interpretation abilities of artificial intelligence compared to humans, as well as insufficient adaptability and missing abilities such as emotional intelligence. In addition, it is emphasized that the lack of adaptability of machine learning causes artificial intelligence to be incompatible with general intelligence in changes. This study highlights the importance and solution power of explainable artificial intelligence for decision support systems as an optimization problem from the perspective of management sciences. Issues such as algorithmic bias and inequality metrics are important considerations for decision-makers in the use of artificial intelligence technologies in management sciences. Furthermore, since the lack of emotional dimension of artificial intelligence may lead to algorithmic biases, it is important to consider emotional intelligence components in the development of these technologies. This study aims to contribute to the societal benefit of the use of artificial intelligence by also drawing attention to its ethical, legal, and societal dimensions.