Application of Generative AI (ChatGPT as example) in Risk Management

This article examines the extent to which models based on artificial intelligence are used in banking services (more specifically GPT - Generative Pre-trained Transformer - models), and in risk management units in particular. It also assesses the quantitative capabilities of ChatGPT through the analysis of two business cases, one focusing on the broad aspects of risk management and the other on European call pricing. For the first use case, we compared the obtained results with another paper to identify if there is any improvement in the used model. The paper highlights also the need for futher analysis of ChatGPT’s generated answers to overcome the tool’s limitation when it comes to quantitative questions in particular.