Abstract

Organizations want to leverage AI systems by effective use of their data and reduction of risk espoused throughout the operation of AI systems. Given the aim for risk reduction, we see the need for highlighting several AI governance areas. In the paper we explore risks and limits connected with the application of AI models in financial modeling and its application in practice. We will present approaches to overcome limitations and potential biases. Based on an example model, we highlight certain issues and problems. We review current policies and present their requirements. Finally, we discuss how many policies are aligned with the technical possibilities and limits and based on that develop suggestions for changes.